Wow, my posts are rather behind on the times. Let's clear the backlog:
December 28th: Sarah arrived! While DFW's arrival-status screen was
mildly defunct (see below), we managed to meet up successfully, and went
home and had tasty chili for dinner. After dinner - long conversation
on my blue couch until late-o-clock.
December 29th: Sarah and I went on a two-mile run in the morning.
She's training for a half-marathon, and I'm trying to get in a habit of
exercising. I ran in my Vibrams, which was fun, but resulted in hotspots
on my feet and extremely sore calves the following three days. She also
led me through a decently long session of yoga afterward to stretch out.
I can't imagine how much worse it would have been if we hadn't stretched.
We started on a 500-piece glow-in-the-dark puzzle of a (rather creepy-looking)
fairy or goddess or something standing with a long flowing cloth on a wave.
We also watched an episode or two of the Dr. Oz show with my mom. It's
quite simplified from what I'm sure Sarah studies in Nutrition, but the
intent is to reach a broad audience with factual, actionable information,
which it does. We stayed up late again, working on the puzzle and talking.
I've quite missed having my close friends in physical proximity. I can't
expect to have the same level intimacy with my new friends in California yet,
but I find it much harder to build such relationships without living in the
same building as them. This will require further effort on my part.
December 30th: Last day with Sarah. Joni and Juliet (two of my mom's friends)
came over and shot the breeze for a bit. They're entertaining folks.
That afternoon/evening, Sarah and I drove down to Mexia, where we joined
four other dear friends from A&M for dinner and an evening together. It
was a really special treat to see everyone. I played
the Speights' piano, joined in a round of 42 (which my team won),
played a match of pool (which my team lost), and caught up on the lives of
my friends after our diaspora. We've had varying levels of struggle with
our new lives, but we're all coping and making progress. It's a beautiful
thing to have friends that you know will understand a joke you make, or
that think of the same memory as you from some conversation. That's another
thing I'll have to build in Berkeley.
As the evening progressed, more friends parted ways, and eventually I left
Sarah with Lara and her family and drove back to Dallas.
December 31st: After spending a week seeing people, I needed a day
off from socialization. I spent most of the day relaxing at home (and
moaning about my poor aching calves). I beat my dad at a game of chess -
I managed to win a material somewhat early on, then exchanged down until
I was a bishop and three passed pawns up, at which point my dad resigned.
I've no real gauge of my skill level, but I like to think that I'd beat
most people who don't play chess regularly, and would lose to most people
that do. Whatever ranking that is.
I finally managed to get a copy of the Console Hacking 2010 presentation slides
and video from 27c3, which was absolutely fantastic. Even
Penny-Arcade wrote about how awesome
it was. Basically: the PS3 is completely, utterly, permanently hacked. Sony
screwed up their cryptography implementation in such a major way that their
secret signing key could be reconstructed from the signatures themselves,
distributed with all software they've signed so far. Oops. Perhaps the best
part is that Sony can't reasonably change the signing key, or all software
released for the PS3 will stop working, so the hack is reasonably permanent.
January 1: I went to lunch at Hong Kong Royal
Restaurant (dim sum!) with Matt and Katy. It's
quite good - 4 star rating on Yelp, and our experience was also positive.
After lunch, we went to Fry's Electronics and wandered around for a
bit. I found nothing particularly enthralling this trip. I definitely
wouldn't mind having a nice Weller soldering iron, but I don't fancy
trying to pack one in a carry-on suitcase. I was surprised to see the
Pragmatic Bookshelf books at Fry's,
but I guess that makes sense after all.
After Fry's, we stopped by Target seeking a season of NCIS on sale
for Katy's mom. We didn't find it, but I did buy the Tron: Legacy soundtrack
by Daft Punk, which is pretty great.
My family met Nicole and her sister and mom at
Osaka, a sushi buffet,
for a dinner of epic proportions. I ate my fill and then was SUPER SLEEPY
(yay, food comas!). So much delicious food.
Following dinner, we all returned to our house to open xmas (new year's?) gifts.
I got a slick black hoodie that read "Powered by Open Source" from Ariana (it's so me!),
a neck pillow for plane flights from Elaine (got a couple of long trips coming up),
and a copy of Cooking for Geeks from
Nicole (so I can really know what I'm doing in a kitchen!). I had a good time, and
it was nice to relax with their family again.
January 2: My friend Thomas from TAMS and I were supposed to meet
at the movies to see Tangled, but it turns out that he was hit by the
iPhone alarm bug
(yes, the second one - there was also one around
daylight savings time.
Why Apple hasn't fully audited that code yet, I have NO idea.) so I saw the
movie with my family, and we met up afterward for cards, chat, and video games.
He's just finished a semester abroad and his undergrad studies; so he'll be
applying to both jobs and grad schools over the next few months.
Tangled was cute. I can't help but analyze the CGI of every movie I see,
now that I've taken computer graphics. Yay,
subdivision surfaces!
Now that I've got
twofriends
messing around with Shortlog, my minimalist blogging engine (crazy, I know),
I've come to the realization that I need to make it easier for people
to access/use my software. I need to make it less tied to my particular setup,
more easily configured, and better-documented.
Then I realized that this applies not only to shortlog, but the rest of my projects
as well. And the first step in using someone's code is obtaining it.
So here you go. I've gone and organized a few
things enough to get a decent-looking gitweb page up to make the source available.
Documentation and modularity will come in due time.
Another interesting thing I noted today: most of my computer usage is
a giant hierarchical organization of data segments. Multiple desktops
play home to multiple tiled windows. Most of these windows are either:
web browser windows, each with a multitude of tabs, or
terminal windows, each with a multitude of tabs.
While there are also occassionally chat windows (with several tabs)
or the one-off LibreOffice window or file manager, this is my most common
interaction with my computer - finding and selecting the desired context to interact with.
This sounded remarkably like Ben Schneiderman's Direct Manipulation
Interfaces. To
make context switches cost less, I (subconsciously) put likely neighbor
contexts next to each other. I may set rules, like "desktop 1 is for
communications, desktop 2 is for OpenKinect development." The catch here is
that some end nodes (application windows) fit into multiple places in the
heirarchy.
My question is then: is there a way to keep the benefit of being able to keep
such rules for quickly finding the relevant context and still preserve relative
context nearness?
OpenKinect has synced master and unstable, which means I need to get my behind into gear
and get my API changes in before the upcoming 0.1 release. Stuff is still segfaulting in
libusb, and I have no idea why. Crap.
Kyle and I have been talking about features we'd
like to add to Shortlog, the engine that powers this blog. Things we've come
up with so far:
Dokuwiki-inspired syntax. I googled
about for a bit, and it seems that no one else on the internet has bothered
to reimplement the dokuwiki text formatter in another language yet. This
will be the third parser I have written, and in a different language to
boot.
Optional post titles
Modular internals
A config file
Stylesheets (Kyle has already made progress toward this end)
If you dear readers have any suggestions for things I should add to my blog,
I'd love to hear them! Leave a comment!
My calves are sore as all get out, thanks to standing up on my tiptoes 242
times (it was a competition with my mom, for the record. I won.). Hopefully
I'm gaining muscle!
Okay, I've done a really poor job of keeping my blog up-to-date recently.
Apologies, dear readers. Let's clear the backlog:
January 7: I met up with Sarah Hodde and Eric Rasche as they were passing
through Dallas. We went to a Panera Bread. I'd never been, but Eric was an
experienced Panera client. I love paninis. We talked for a good few hours,
catching up with the happenings of each other's lives. It was so good to see
both of them - apparently, we've all had really rough semesters.
January 8: Eddie Eigenbrodt and Julia White, two awesome folks from
A&M (that I had sadly not seen in quite some time) were engaged.
Today was their wedding!
I drove down to Dripping Springs in the morning and found the event center
at which the wedding was set to take place. It was supposed to rain,
and the wedding was to take place outdoors, but the sun peeked out from
behind the threatening clouds just in time for the ceremony to begin.
I didn't know which side of the aisle to sit on - I met both Eddie and
Julia my first semester at A&M, and had several classes with them both.
Everything was beautifully decorated in black, white, and red. Julia has
always loved penguins, so they made numerous appearances in the decor,
including a pair of plastic penguin figurines (one tall, one short) sitting
atop the wedding cake. The dinner tables were numbered, but not sequentially.
Rather, each table had a number that was significant to Eddie and/or Julia in
some way, like pi, Avogadro's number, Planck's constant, 235 (for Uranium 235,
since Julia is a nuclear engineer), or 1050 (the muzzle speed of a .22 pistol
in feet/sec(both Eddie and Julia were involved in pistol team)). I got to sit
with the pistol team at table 1050, and they were interesting people. In a
"wow, small world!" moment, I met the brother of Andy Echols, whom I knew from
the A&M computer society. In another "wow, really small world!" moment,
Mark Browning introduced me to Kate Stewart, the Ubuntu release manager. Turns
out she's friends with Julia's dad. We chatted for a bit about the open-source
community and contributing and how cool she finds it to see the project come
together from the top, but also run into the individual hackers that make it
happen from time to time and hear about the pieces. Pretty awesome
coincidence.
All together, the wedding was lovely and perfect and I'm overwhelmingly happy
for Eddie and Julia. We sent them away by the light of a hundred sparklers,
and it seemed like everyone had a splendid and joyous time. And inspired by Gabe,
I lit two more sparklers and used them as poi. Another wedding guest said I
"looked like I knew what I was doing" which was awesome because I totally didn't.
ANYWAYS: Congratulations, Eddie and Julia! I wish you all the best in your life
together!
January 9: The night of the wedding (after I'd left and checked into my motel,
thankfully), I came down with some bothersome sinus bug that resulted in my
blowing my nose nonstop for practically the entire night. I coulden't breath
out of either nostril, and had an awful time falling asleep. Around 4:30, I
finally managed to stay asleep for more than an hour, eventually leaving to
drive back up to Dallas around 11am. Oh, it was snowing. There was snow and a
small amount of ice on I-35, which had a negative impact on visability, highway
speeds, and my level of calm. I made it home about 15:30, at which point I
finally got to take some proper decongestants and nap and eat. The wedding the
day before was so awesome that it was totally worth it though.
January 10: Didn't accompish much of anything. Tried going for a run, but it was
ridiculously cold and I was done running after about 10 minutes. I covered probably
¾ths of a mile. Gah.
On the upside, the frigid air and running cleared my previously-blocked nostrils
out rather well.
January 11: See January 10. Although I did learn about how NOT to use libusb
from different threads. Which is relevant since I still don't understand why
my high-res patchset for libfreenect crashes.
I got a bit stressed out at the prospect of packing and leaving, which was
strange, since I've been fine doing so for the past 5 years or so now.
January 12 or, "When the **** hits the fan": I was scheduled to fly to Boston
today. Through Atlanta. For historical value, I'll note that Boston is
currently having a massive blizzard (30mph winds, over a foot of snow), so no
flights got in today. Furthermore, Atlanta is on the tail end of
uncharacteristically cold weather, including so much snow that their roads iced
over. Atlanta owns a grand total of 8 snowplows for the entire city. CNN did
a viewer-submitted piece today showing a guy ice skating down Peachtree Street,
Atlanta's main street. So Atlanta has not been terribly on top of flights
either.
But I was hopeful. Neither of my flights were cancelled when I woke up. I
hugged my dad and sister goodbye as they left, and then my mom drove me to the
airport. Both flights were still a go. I went through security. About 10
minutes before my first flight boards, my dad calls me, letting me know that in
the past minute, my second-leg flight had been cancelled. Along with every
other flight from Atlanta into Boston except one that arrived just before
midnight. I quickly went up to the service desk, and the helpful woman there
managed to get me one of the rapidly vanishing seats on that last uncancelled
flight.
My DFW -> ATL leg went smoothly. When I arrived, the second-leg flight was
still a go. It remained that way for about three hours, and then I got another
phone call from my dad, again helpfully letting me know that my flight had just
been cancelled and that I should seek another. (How he managed to get this
info so quickly is beyond me, since he got me the word faster than the
Departures screen at the airport. I have this picture in my mind of him
refreshing a flight status webpage every 10 seconds all day long, which makes
me laugh.) So I hauled between terminals to find the Airtran Customer Service
desk, and they put me on the first flight out on the 13th. We'll see how that
turns out tomorrow. For now, though, I've pulled up a comfy patch of carpet
in a corner near a power outlet for the night.
Oh, one other thing about Atlanta: the wifi sucks. Well, technically, not the
wifi - the access points ping just fine, but they block access to the general
web with what's probably a massively overloaded squid proxy. That might
explain why the redirecting proxy managed to drop approximately every other
connection I tried to make.
"But!" you might say, "don't you have your Nokia N900, with which you can
tether at delicious speeds?" I do indeed, but it turns out that T-Mobile's 3G
backend was also having issues. I'd get an IP, but couldn't ping anything at
all. Eventually I figured out that I could force the phone to use only EDGE
instead of 3G bands, and that worked. That wasn't until much later in the
evening, though, so most of the time I was trying to check my flight status
with either of two failing paths to the Internet.
BUT now I have EDGE so I can write this lengthy and belated post. With any
luck, in twelve hours, I'll be in Boston.
A series of naps on a matched pair of seats provided me with enough rest to be
surprisingly cheerful this morning. My flight switched gates (twice), but it
looks like everything will be on time now.
I met a woman and her daughter who were also travelling to Boston and had been
stuck in Atlanta for two days. The mother did computer science in the 80s with
CPM, and the daughter was planning to study math and physics. We chatted for a
bit about technology, politics, and the future of our nation.
UPDATE: I have arrived in Boston. My aunt picked me up, and I am now settling
down for a nap.
This post is backdated, on the grounds that I should have written it on the
17th and I have a decently large post for the 18th to write as well, and that
they're kinda disjoint topics. Some people might say that my blog engine
should support multiple posts per day, and they'd be right (it is, in fact, on
my roadmap), but that feature is not at the top of my priority list right now,
and no features have gotten much attention lately, so they can deal. So,
without further ado: my Boston trip.
Just because it deserves a paragraph of its own for emphasis: I've had an
absolutely fantastically awesome time the past few days.
As previously mentioned, I
successfully made it up to Boston. That night, my aunt and I went and visited
my cousin Erik, his wife Julie, and their two children and two cats. Julie
made a completely delicious chicken pot pie from scratch, which I had several
helpings of. Samuel is now able to eat corn, which is exciting as both of the
kids have the strangest mix of mutually-exclusive allergies. The kids and I
played with LEGOs and rolled a ball, and later that evening I explained how
objects fall at the same speed (up to air resistance), complete with
experiments involving dropping all sorts of things from arm's length. Julie
read excerpts from the Maine Dictionary
with the proper accent, and we all laughed and had a merry time.
The next morning, I packed up my effects and accompanied my aunt to a doctor's
appointment. It took quite a while longer than she had expected. Fortunately,
we arrived at MIT just in time for me to drop off my effects in Death From
Above (the Mystery Hunt team I joined) HQ before running over to Lobby 7 for
the kickoff event. And oh, what a kickoff. When I arrived, there was a string
group playing some procession theme that I couldn't quite place. Lobby 7 was,
of course, packed with people, and I managed to nearly walk into Nicole before
recognizing her. On the second floor, there was a man dressed up as a
preacher. Slowly, the music turned from "generic wedding procession" into
"wedding adaptation of the Mario theme," and Mario and Peach arrived on the
scene. A hilarious sequence ensued involving Bowser crashing the wedding
before the couple could be joined in "holy mushroom-ony" and kidnapping Peach.
Hunters would have to help Mario rescue Peach by solving puzzles and defeating
Bowser. The hunt began shortly thereafter.
I took a break for a bit to go meet with Michael Bernstein and the other
members of the User Interface Design Group.
I had the pleasure of joining their 1pm group meeting, where I shared about my
work with the Kinect, listened in on their preparations for the CHI 2011
Workshop on Crowdsourcing and Human Computation,
and learned about some of the their current focuses.
After that, the next 48 hours or so were dedicated fairly strictly to the Hunt.
There's no way I'd be able to do the hunt justice in a blog post, so I shall
simply link to the Hunt website
and discuss a few of my favorite puzzles. Before I begin, I should note that
the hunt had five rounds, each focusing on a different video game - Mario, Mega Man,
Zelda, Civilization, and Katamari Damacy. A few puzzles that I either helped solve
or just found interesting:
The Mario World 1-2 meta
(which I figured out while in the bathroom, for what it's worth), followed by immediately
backsolving
two other puzzles from that round.
N-Tris, a puzzle in
which you played two simultaneous, interacting (thankfully not real-time)
tetris games with the option of kicking a piece to the other drop column,
eventually constructing a QR code.
Good Vibrations,
an audiopuzzle in which the left ear plays snippets from 16 different songs
and the right ear plays "Good Vibrations" by the Beach Boys. Turns out
each song title heard in the left ear matches up with a vibrator sold on
the GoodVibes.com (NSFW!) sex toy store.
[You Shall Understand What Hath Befallen] (http://ihavetofindpeach.com/puzzles/mega_man/you_shall_understand_what_hath_befallen/)
, a puzzle where you reconstructed the moves made in Reversi games, noted the forking plays, and interpreted them as semaphore.
Timbales, a puzzle involving Latin phrases and (rather amusing) fake definitions.
Powder Monkey, a
puzzle in which you received 18 baggies of various white powders.
Obviously, to solve the puzzle, you'd need to identify them. When we asked
HQ if they were safe to ingest, we were told that "none of the powders
should cause lasting damage to a person of average constitution." One of
the other teams apparently wound up insufflating the cement.
One night of the hunt, my friend Lora from Berkeley IMed me and thanks to Death
From Above's awesome organization, it was really easy for her to join in the
hunt. She donated her expertise on France to help solve
Pointillisme.
The more brains, the better!
Another interesting event worth mentioning was when Mark Browning called me.
He needed to get to some documents on his computer proving marriage and had
only HTTP access. So he called me up and had me SSH into his desktop and
symlink his documents folder into his public_html folder so he could find the
documents he needed to sell his car. I was happy that I could help my friend
in a time of need.
Sleep during Hunt was minimal. I took a couple few-hour naps on the couches in
CSAIL (they're quite comfy!) to refresh myself when I started hitting a wall.
The rest of the time, I was running on pure adrenaline.
After Hunt ended, I helped clean up CSAIL and vaccuum the rooms we used. I met
Bill Reading (who was on the winning team last
year, and thus helped run the Hunt this year) at the Hunt wrapup session in
26-100, and he shared all about his trip to Germany for 27c3, which sounded
awesome.
And after that, I went back to East Campus, where I dropped off my effects, and
then joined a dozen or so friends at the Cambridge Brewing Co. for dinner,
where I had an excellent BBQ burger. Upon return, I met up with more old
friends, made some new ones, and broke out my Kinect. I got to observe MIT
kids making fools of themselves in front of the TV. It was fun and
entertaining to watch. Later, we watched How To Train Your Dragon and Wall
Street 2, exchanged backrubs, and played Super Smash Bros. Brawl. I took a nap
around 8am after a nice girl named Cheetiri made me breakfast. I had meant to
visit the 6.270 lab (LEGO robots!) but failed to do so. Maybe next year.
A touching moment: as I was preparing to leave for the airport and giving my
goodbye hugs, my friends from EC exclaimed that I wasn't allowed to leave.
Even though I'd only shared parts of a single semester with them, they
appreciated me and welcomed me warmly, and told me I had to come back and visit
again. I was surprised. They all treated me like I was a real MIT alum, which
made me feel really special.
I took the T back to the airport. I couldn't remember if I had a CharlieCard
in my wallet, so I shoved the billfold up against the reader, and it worked.
Guess I still carry one, though I've no clue where.
My flight from BOS to SFO was uneventful. My line at security only had the
metal detectors, so I didn't even get a chance to make a statement about the
backscatter machines by opting out. Oh well. I moved with purpose once we
arrived in SFO and BARELY made the 10:27 BART train out. I arrived home at
about 20 minutes before midnight, and found a box on my bed. Assuming it to be
Christmas gifts that my parents shipped, I left it alone, took a much-needed
shower, ate some dinner, and fell asleep.
Woke up about an hour before my appointment with my counselor. I decided to look at
the box that was sitting on my bed when I arrived home last night.
Turns out that the box I mentioned yesterday was neither xmas presents nor from
my parents. It was also not from any address I recognized - some Brian
Jeffries guy in Kentucky. Curious, I opened the box. Inside, I
found...another box. Somewhat heavy. This one had a handle, and a rather
unique design on the outside of an exploded-view of rocket of some sort powered
by a mouse wheel. It seemed strangely familiar, but I couldn't quite place it.
Then there was a sticker on the back of the box that threw me off - a
lithium-ion battery warning.
"What's going on here?" I thought to myself. "I don't recall ordering anything
recently, and I don't know anyone who lives in Kentucky." Confused, I applied
my skeletool blade to the seal on the box. I opened the lid. And then I saw
this piece of paper:
"Oh. I know what that is!" thought one part of my brain. "No way!" went
another. "Wow, that was a while ago!" offered a third. A fourth added, "Good
thing you didn't open the box last night."
For those of you who still have no clue what I received: my box contained a
Cr-48 ChromeOS laptop,
courtesy of Google. I had applied to the
Cr-48 Pilot Program
back in early December, but figured my chances of receiving a unit were quite a
long shot, and completely forgot about it over break while I was in various
other cities.
Looks like I was mistaken. A second review of the box indicated that it was
delivered on December 21st, which meant that it sat in my apartment complex for
a good couple weeks before my housemate could bring it in. I'm pleased - this
suggests that my apartment complex and all who pass through it are trustworthy.
Obviously, I did what any self-respecting nerd would do: I tore off the rest of
the packaging, popped in the battery, and powered it up. It booted very
quickly, as advertised - about 10 seconds. Logging in and getting started was
easy - the hardest part was getting my just-out-of-bed hair into a condition
suitable for a login picture. Bookmarks and the like synced from the Google
servers, and I was left with a Chrome instance not unlike those on kraken and
phoenix, my desktop and laptop, respectively.
And shortly thereafter, I realized I had about 15 minutes to eat breakfast and
get my rear end over to campus, or I'd be late for my appointment. I set the
Cr-48 on its charger (also a delightfully small object) and ran to campus.
I analyzed my present state of being and plans for the semester with my
counselor, and was satisfied with the insights obtained. Afterward, I dropped
by the BiD lab to visit my friends. They described me as looking much
healthier and happier. They're completely right. The past month has done
amazing things for my mental health. I've gained back the weight I lost last
semester.
I returned home so I could buy groceries (since I still had very little in the
house) and played around with ChromeOS. My thoughts so far:
The keyboard feels very nice. It'll be interesting to see how the
rubberized keys wear over time, as well as how they handle sweaty hands and
so forth.
The touchpad drives me bonkers. It's not quite as flakey as the one
present in the circa 1996 lapbrick that I used back in the day, but it
tended to jump around a lot and mess up scrolling pretty often. I may
start using an external mouse. Alternately, I'll start using the keyboard
for scrolling wherever possible.
I miss focus-follows-mouse. I've gotten so used to throwing the cursor
around without having to click that having to click on things feels
annoying.
I do not miss the weight nor the heat that the Cr-48 doesn't provide my
lap.
Suspend and resume are incredibly fast. Faster than you can reasonably
safely close and open the lid. Reconnecting to the wireless network, on
the other hand, isn't quite so snappy, although for that I blame my router -
if I were running hostapd on my desktop, it'd probably be equally quick.
I lasted for about three hours before I decided that I needed a terminal
(which is kinda sad, really) so I enabled developer mode by flipping the
hidden switch. It might be worthwhile to look for an SSH-like webapp to
deploy on my server. I wouldn't feel comfortable securing it with anything
but client certificates, though, so I'll have to look up if ChromeOS
supports them or not.
Speakers are not completely awful, but nothing wonderful either. As would
be expected for a 12" laptop.
This is whiny, but I wish I could figure out how to paste into aterm.
Probably could read some documentation to figure that out.
This is also whiny (and probably way outside Google's target audience) but
I also wish I could get ssh to read my user config file and private keys
without me having to specify them with -i and -F.
Last whiny thing: where did vi go? It's not in my $PATH and I can't find
it (although there is some locale stuff still present). Nor is there any
trace of nano or pico. Weird. Good thing there's qemacs; I was beginning
to worry I didn't have a text editor (besides cat and sed).
Nothing has crashed yet, as far as I can tell. This bodes quite well, but
makes me feel a tad less useful as a beta tester.
I've yet to try out the free 100MB/month of Verizon 3G, or plugging in a memory
card with photos, or in fact any USB devices. I'm sure the opportunities for
each of these will arise as time goes on. This notebook has much better
battery life than my Dell Latitude D830 (although it's much less powerful) and
is definitely more convenient to break out at a moment's notice. Thanks,
Google. This was a super-awesome welcome-home present. I'll do my best to
break your software so others won't have to.
That was today's round 1 of awesomeness. Round 2 was my housemate getting a 10
cup fuzzy logic rice cooker. You can set a time for this thing so your rice
will be ready when you get home. It will keep rice moist for a full day. It
plays a happy tune when the rice is ready. Andrew and I tested it out tonight,
and the brown rice was pretty awesome. My rice consumption is going to at
least double this semester.
I'm curious as to what's happened in the past five months, with regards to the
Skype protocol. I know that in July, a talented reverse engineer did what
those before him
didn't want to bother with
(PDF, slide 52/72) and reimplemented the seed-to-RC4-key function (the core of
Skype obfuscation/encrypted messages) in nearly 3000 lines of C. He was
supposed to present at 27c3, but apparently didn't. The code, however, is
still out there.
As a major fan of open-source software, I'm wondering: has anyone bothered
documenting the (current) Skype protocol? I'd love to see a free/open-source
implementation of the Skype protocol, say, for Telepathy. If the docs exist
out there somewhere, then it shouldn't be terribly hard to write - Telepathy
was made to be quite flexible. If, on the other hand, the protocol is still
undocumented, well, maybe I have a good semester project. How to construe this
as a computer security project...hrm.
Switching topics: the GNOME 3 website has
launched. I can't help but read down the feature list and think "wow,
KDE's had that for X releases!"
Activities? Plasma has had these for years. Notifications system? KDE led
the design of the freedesktop.org spec that's in use by all. And it ships in
4.5. Keyboard launch/switch? See KRunner and its multitude of searchable
capabilities. System settings? Let's not go there
again. Window
tiling? Shipped in KDE 4.5; now I can't operate efficiently without it.
Netbook interface? Marco Martin did that in time for KDE 4.4. Dark
application themes? Is that even a feature? (If it is, KDE's had the Obsidian
Coast theme since at least 2006.)
I'm not saying that GNOME isn't an awesome software project; it absolutely is,
and provides a ton of people with exactly what they want and need. I'm just
noting that KDE is the innovation powerhouse and ships good ideas, despite
its comparable lack of commercial funding.
Luke's birthday is coming up, so he, myself, and a couple more friends are
going on a 7-mile hike in Mt Diablo State Park. I'm going to wear my Vibrams
(and pack my socks and hiking boots, just in case).
Today I received my first spam comment. Yay, my blog is moving up in the
world!
Update: The hike was excellent, although we wound up not going on the full
7-mile loop, we did go up to the summit and did a brief free-climb on a rock.
It was about 15°C out, but was quite windy, particularly at the summit. I
enjoyed wearing the toe shoes on the way uphill and for the rock climbing, but
changed into my boots for the descent - too many sharp rocks, steep hills, and
too much loose gravel for comfort. Here's a few pictures from the venture:
Awesome view.
You could see a lot of counties.
The sky was gorgeous too.
A rock formation we happened upon. And happened to climb upon.
The birthday boy, on top of the world.
More of our merry hiking party atop the rock.
Me, with ridiculously wind-whipped hair.
Andrew has a picture on his camera of all six of us, which I will post once I've obtained it.
I woke up this morning in tears. I haven't had a nightmare like that in a very
long time. On the upside: I'm sleeping well enough to dream!
The EECS Career fair was today. When I told the Microsoft recruiter that I had
worked on the open-source Kinect drivers, he wrote a large E on my resume and
circled it. I believe this to be a good sign. Other fun from the career fair
included a rather bubbly Autodesk recruiter, taking a brief C quiz for nVidia,
having the Google recruiter recognize me and remember that I'm a PhD student and
who my advisor is, and getting a scarf from VMware. I won a keychain with 64GB
of unusable flash (shiny, though!) at the Intel infosession this evening.
I made a cucumber/bell pepper/tomato/ricotta salad for dinner. It's been a
while since I broke out the balsamic vinegar, but it turned out well.
Two sequential names on my Jabber contacts list (both starting with J) have
statuses that consist of the same word four times. The two people in question
also don't know each other, which makes this even more amusing. For the curious,
the words are "DNS" and "sigcomm."
I had my first meeting with my advisor since I returned to Berkeley. I was
quite nervous, particularly since I hadn't made any progress on projects since
leaving town, and my subconscious attitude toward research had worsened. In
three minutes, he managed to eliminate all of the negativity and frustration
and worries that I had about classes and research and everything else. The
next twenty consisted of planning who we'd contact for research plans, summer
opportunities, and so forth. I left excited and feeling good.
UPDATE: I managed to fail to notice that the State of the Union Address
that follows (and which I offer my extensive critique of) was the 2010 State of
the Union, not the 2011 State of the Union. How embarrassing. Fortunately,
most of the criticism still stands.
This evening, President Obama gave his State of the Union address. If you're
an American and haven't watched/read it yet, please do so
now.
Besides, if you don't, you'll probably be a little confused by the rest of this
blog post, because I'm going to go through some of the things that stood out to
me. Feel free to follow along in the
transcript.
Dad, I expect you to share your thoughts. I know you'll have some good
insights; you always do.
"Next, we can put Americans to work today building the infrastructure of
tomorrow." INDEED. America has done a ridiculously poor job of investing in
infrastructure, as compared to China, Japan, and Europe. I wish we had better
rail systems in more places. Whether Internet or phone service should be
considered public infrastructure or not raises some interesting questions.
On clean energy: the success or failure of this plan will depend largely on how
such clean energy systems are subsidized. On principle, I'm all for clean
energy, and renewable energy. I believe it's the only viable solution for the
very long term, and that eventually, energy will become the primary resource
(and even currency) of the world. Unfortunately, we're not at a point where
clean energy technology is efficient enough to recoup its own costs over its
lifetime, and we do a great disservice by subsidizing production of clean
energy (think commercial power plants) rather than innovation and research to
make this technology more cost-effective. See this excellent Forbes
article
for some more background. The picture the data paints right now is that due to
government subsidies, clean tech deploys ineffective technologies that do
not recoup their costs, and since the companies are profitable (because of
the subsidy) they do not bother investing in further research. So my
takeaway: invest tons in clean energy research, not deployment. When it's
good enough to be viable on its own, feel free to subsidize national
deployment to promote adoption of the viable tech.
On exports: well, duh. The way you wind up paying off government debt is
by selling things that other people want. And the way to do that is to create
value. Note: intellectual property only counts as value when you're doing
business with copmanies that follow your rules. Guess which developing
economies aren't interested in following your IP laws? So yes, we need to up
our exports.
"In the 21st century, the best anti-poverty program around is a world-class
education." Indeed. The Jews figured this out centuries/millenia ago; they've
always valued education, and that's why they've done rather well financially
throughout history. Unfortunately, as a result of their success, they've
aroused the anger and vengence of many jealous, less forward-thinking
cultures.
"This year, we will step up refinancing so that homeowners can move into
more affordable mortgages." I'm not sure I like this one so much. While I'm
not interested in denying people a place to live, it's my understanding that a
major factor leading up to the housing crisis was that people buying houses
could get WAY more credit (and thus, buy way more house) than they should
reasonably have been afforded. With the unfortunate American attitude of
entitlement and living a luxurious life, these folks wound up buying more
expensive houses than they could really afford. The result was foreclosures
and the housing bubble bursting. While offering refinancing might be the
compassionate thing to do, part of me thinks "you shouldn't have bought that
much house, and you're getting what you deserve." Government offering to fix
all ills produces a whiny, expectant citizenry, rather than a responsible
one.
And on that note, healthcare. I don't know enough about the healthcare system
to be a fair judge of most factors. However, this statement stuck out at me:
"Patients will be denied the care they need. Small business owners will
continue to drop coverage altogether. I will not walk away from these
Americans, and neither should the people in this chamber." I take some amount
of issue with this - the US spends more on healthcare per capita than every
other country I could find data for. (Source: EOCD Health Data
2010.)
It seems odd, then, that given that we don't have nationalized healthcare
(well, not really deployed in full yet) that we should be denying so many the
"care they need." Ultimately, we have to put a price on how much we're willing
to spend on a person - that's what insurance is, after all. If we can't afford
to pay as much as we want to, then in the long run, yes, patients must be
denied the care they need. It's not sustainable for a nation to do otherwise.
I recongnize that the insurance system is inefficient and sometimes
exploitative. However, the big picture facts remain the same - we can't afford
to spend this much per capita on healthcare, particularly not by government
mandate. Take a look at some of the European countries that are falling apart
because they can't afford to live the way they want to. Another problem with
universal healthcare is that once it's granted, there's really no way to
reverse that decision, should it be economically infeasible.
Which brings us to the question of the deficit. This is where I shat a brick,
to put it simply. "Now, if we had taken office in ordinary times, I would have
liked nothing more than to start bringing down the deficit. But we took office
amid a crisis." It's remarkably convenient how often there's a crisis in
Washington. Or a war. On a poorly-defined concept that won't go away. Like
the war on drugs. Or the war on terror. A hallmark of an authoritarian state
is that it is in a constant state of emergency. I'm not saying we are under a
police state, but I think that "crisis" is an excuse. But I digress.
"Starting in 2011, we are prepared to freeze government spending for three years." Excellent! But wait, there's more.
"Spending related to our national security, Medicare, Medicaid, and Social Security will not be affected. But all other discretionary government programs will."
What. The. <censored>. There can be no sacred cows in the budget.
Let's examine the data, shall we? There's a fantastic infographic called
Death and Taxes that gives a
wonderful, easy-to-follow graphical presentation of the US government spending.
While the main item in the graphic is the Federal Discretionary Budget, if you
look to the far right, you'll see the pie-graph penny that represents the Total Budget.
We're spending a total of 3.834 trillion dollars, and we're slated to collect
2.567 trillion. While this alone is alarming, let's cross-reference the four
items of spending we said would be immune to this government spending freeze.
By the way, "spending freeze" just means no adding new expenditures. We'd just
keep all our current spending levels.
National Security spending: 895 billion. Medicare? 491 billion. Medicaid?
297 billion. Social Security? 730 billion. That's 2.413 trillion dollars.
That's 63% of the federal budget that we just said would be immune to a
spending freeze (not even cuts!). And those four items alone are 94% of our
receipts (income to the budget). Throw in the national debt interest (251
billion, which we can't exactly elect to stop paying), and you exceed 100% of
our receipts. Boy, it's going to be awful hard to get spending in line with
income when you're both cutting taxes and refuse to even freeze spending on
costs that amount to greater than your entire income. Again: what the
****. This is not an adequate plan for restoring solvency to our nation. The
whole speech lost credibility in three sentences.
"We've made substantial investments in our homeland security and disrupted
plots that threatened to take American lives." Oh, we've made investments, all
right, but said investments have hardly disrupted any "plots that threatened to
take American lives." Increased public vigilance and the attitude that plane
hijackers are not just going to detour you to Cuba are what has stopped
would-be terrorists. I've
complainedbefore about how
the TSA doesn't actually provide real security, just
security theater. On
the other hand, they've increased government spending, which IS a major issue.
And don't say "but this creates jobs for Americans!" because I will smack you
down with the broken window fallacy.
"I'm also calling on Congress to continue down the path of earmark reform."
Hooray. Can't happen soon enough. "But we can't stop there. It's time to
require lobbyists to disclose each contact they make on behalf of a client with
my administration or with Congress. It's time to put strict limits on the
contributions that lobbyists give to candidates for federal office." Great.
I'll believe it when I see it - neither party is exactly incented to limit
their income from lobbyists, in kinda the same way that neither party is
incented to make it easier for third parties to acheive meaningful
representation.
There was a bit on how Iran is seeking nuclear weapons, and I couldn't help but
think of Stuxnet at this quote: "That's why the international community is more
united, and the Islamic Republic of Iran is more isolated. And as Iran's
leaders continue to ignore their obligations, there should be no doubt: They,
too, will face growing consequences. That is a promise." It does seem likely
that it was the U.S. and Israel behind that
worm.
I'd cite the NYTimes article directly, but they've decided to hide all their
content behind a paywall. Pity.
State of the Union: some things good, some things misguided, some big promises
that I'm skeptical about, and some that weren't big enough.
What are your thoughts? In particular, I'd love to hear where you disagree
with my views and why, so we can all get a broader perspective on the tough
issues of the day.
(Fun fact: this is the longest entry I have written, to date.)
So, my epic post from yesterday refers to the 2010 State of the Union, not
the 2011 one.
Epic fail, Drew, epic fail. I'll be reviewing the transcript (but not the
video, since the white house has decided that it'd rather post the video to
youtube than host the .mp4 for download (scratch that, there's a download link
on youtube (the first time I've seen such a thing - apparently they do this for
public domain videos))) (update: 1.3GB .mp4 here)
and hopefully critiquing the more recent words. I'd like to
point out that most of my critique from yesterday still stands.
Education: statement about the Jews having figured it out still holds.
"We need to out-innovate, out-educate, and out-build the rest of the world."
Yeah; pity that we only do the first of those.
"We need to take responsibility for our deficit and reform our government."
Yes! Please, nation, take responsibility. Reform things more efficient and
spend less.
"This is our generation's Sputnik moment. Two years ago, I said that we needed
to reach a level of research and development we haven't seen since the height
of the Space Race. And in a few weeks, I will be sending a budget to Congress
that helps us meet that goal. We'll invest in biomedical research, information
technology, and especially clean energy technology -- (applause) -- an
investment that will strengthen our security, protect our planet, and create
countless new jobs for our people." Hey, there's that clean energy technology
bit again. Same caveats still apply - fund the research until it can pay for
itself over its lifetime. Only then subsidize deployment. All right. Hey,
this next quote also sounds good. "We're telling America's scientists and
engineers that if they assemble teams of the best minds in their fields, and
focus on the hardest problems in clean energy, we'll fund the Apollo projects
of our time." Sounds like good research to me.
"So tonight, I challenge you to join me in setting a new goal: By 2035, 80
percent of America's electricity will come from clean energy sources." Eh.
2035 is kinda far away. We put a man on the moon within a decade; at a glance,
this is not as ambitious as it should be. Someone who knows energy, please
correct me.
On education: "That responsibility begins not in our classrooms, but in our
homes and communities. It's family that first instills the love of learning in
a child. Only parents can make sure the TV is turned off and homework gets
done." Yes. I'd love for the public to care like that. Keep at it.
"And Race to the Top should be the approach we follow this year as we replace
No Child Left Behind with a law that's more flexible and focused on what's best
for our kids." Yes, please! Get rid of No Child Left Behind; it's a program
that fails our best and brightest, who in turn are the ones who create
businesses and jobs and breakthroughs.
"And this year, I ask Congress to go further, and make permanent our tuition
tax credit -- worth $10,000 for four years of college. It's the right thing to
do." Okay, education is my soft spot. You can just about ask for a blank
check for schools, and I'll sign my name on the line. That said, I'm not sure
that tax credits are the right way to implement this, though. Schools, as much
as I love them, are frightfully inefficient beasts, full of bureaucracy and
excessive layers of management (at least, the UC system is). I wonder if we
could afford students the same level of benefit by making the schools more
efficient with their funds. Maybe I just don't understand because I've been
fully funded to attend school all my life.
Another probably controversial question: will a society consisting entirely of
college graduates be sustainable? Or will we have a similar situation as we do
now, except the person ringing up your fast food order will have a BS instead
of a GED? I'd love to see some data or a study suggesting either way; I've
really got no clue. I'd love for everyone to be that well-educated, but I also
need a reality check.
On illegal immigrants: step 1 must be adequately securing our borders. Unless
you have fully secured your borders, offering amnesty is simply incenting more
illegal aliens to get here faster, so they can be part of the next amnesty
grant. Call me cold, but this should apply to the children, too. Children have
always suffered and will always suffer for their parents' indiscretions.
Sadly, this situation is no different.
On infrastructure: yep, same as last year. I'm curious as to how we'll be
doing the bit with private industry. For example, in the Dallas Fort-Worth
area in Texas, there are a number of toll roads. These were funded with
taxpayer dollars, but are privately owned (some, ultimately by overseas
corporations) and will likely never pass into American hands. Should such
crucial infrastructure be owned and operated by foreigners? Should we be
giving these tax credits to foreigners? If the goal is making competitive
infrastructure with the rest of the world, I think the answer here is no.
So again we'll have to be careful how this is implemented. The devil's always
in the details.
"Within 25 years, our goal is to give 80 percent of Americans access to
high-speed rail." Splendid. I was always disappointed that we didn't have
railcars like Japan, China, Korea, and Europe. Those things are sweet.
The next remark was flippant, but hit me like a sack of bricks. "This could
allow you to go places in half the time it takes to travel by car. For some
trips, it will be faster than flying -- without the pat-down." (emphasis
mine). Whoa. Either 1) we're going to need good security for these newfangled
expensive rail systems that we want all these people to use, 2) we don't
actually need the invasive pat-downs from the TSA, or 3) both. In any case,
this remark is completely off-base. I find the fact that the President is
making light of the erosion of my freedom to not be groped when I travel about
the country very alarming. Very alarming indeed. This is not a perk, and it
should not be a motivator for developing new rail systems. I'd like them to
stand on their own merits, thanks.
On Internet access: by all means, fund it, but make sure the public is getting
what it's paying for. That means strong assurances of Net Neutrality, the
cornerstone of the Internet's equalizing power. Net Neutrality has been the de
facto standard for many years, and thanks to the geniuses behind
TCP, it's
in everyone's interests to reduce congestion, and proportional-share allocation
already works. Without Net Neutrality, big companies will be able to edge out
small startups. If it weren't for Net Neutrality, we wouldn't have Google,
Facebook, or Twitter, because they would not have been able to grow and compete
on a net dominated by large anticompetitive companies.
The FCC decision does not provide adequate Net Neutrality provisions for wireless
carriers. Increasingly, our young and poorer demographics use their cell
phones as their only source of Internet access. Failure to provide Net
Neutrality will ultimately prove an unfair disadvantage to these peoples in the
digital world. I do wish Obama had mentioned the importance of Net Neutrality
in promoting innovation - it ties in so well with the rest of his platform.
On simplifying the tax code: great idea. We'll have to see how that plays out.
On regulations: very diplomatically handled.
On health care: everything I said yesterday still stands. "What I'm not
willing to do -- what I'm not willing to do is go back to the days when
insurance companies could deny someone coverage because of a preexisting
condition." Again: someone's gotta pay for it, and we already spend more than
every other nation in the world per capita on healthcare. And whereas the
healthcare bill was presented as reducing our deficit, it has yet to deliver on
that promise.
"Now, the final critical step in winning the future is to make sure we aren't
buried under a mountain of debt." Ah, finally, the deficit question. He
proposes the spending freeze, Gates agrees we can do without some of our
military funding, still only accounts for tiny portion of our expenditures.
"Now, most of the cuts and savings I've proposed only address annual domestic
spending, which represents a little more than 12 percent of our budget. To
make further progress, we have to stop pretending that cutting this kind of
spending alone will be enough. It won't." Damn right it won't. Let's hear
the rest. "...the only way to tackle our deficit is to cut excessive spending
wherever we find it -- in domestic spending, defense spending, health care
spending, and spending through tax breaks and loopholes." Again, agreed.
"To put us on solid ground, we should also find a bipartisan solution to
strengthen Social Security for future generations." Ugh. No. We should get
rid of the program entirely, and stop playing musical chairs with the debt. It
is not the government's place to provide for the elderly. "We must do it
without putting at risk current retirees, the most vulnerable, or people with
disabilities; without slashing benefits for future generations; and without
subjecting Americans' guaranteed retirement income to the whims of the stock
market." Umm...how do you propose we cut costs, again?
On transparency in government: yes, put all that stuff on the Internet. And more.
And stop playing silly games classifying documents that don't need to be classified.
And keep up with the FOIAs. And quit claiming "state secrets" to cover up things
that the citizenry won't agree with. Thanks.
Blah blah blah terr'ism. Shouldn't have been in all these countries to begin
with; it's about time we got out of there.
And the rest was meant to be inspiring, but didn't really work on me.
So, in review: an awful lot of the same things. Education, science, and
development. Some shenanigans about spending and the deficit. And then
closing with inspirational remarks. Sounds like a recipe.
In another delightful instance of "small world, huh?" I found out that Ian
Haken, a friend of mine from TAMS, is also taking the same Computer Security
course as I am. I didn't even know he was at Berkeley. The conversation went
something like this:
Dr. Song: "Okay, let's take a short break!"
Me: *walks over* "Okay, this is gonna be really weird, but...you look frighteningly familiar. Are you Ian Haken?"
Ian: "...yes, I am." *suspicious eyes*
Me: "Drew Fisher, TAMS class of '06."
Ian: *pauses* "You've got a lot more hair than you used to."
And then after class we caught up with what we were both doing with our lives,
talked about Nokia tablets (he has an N810!), and I gave him a tour of the BiD
lab. We might work on a security project together; who knows.
I'm excited, because GMail deployed HTML5 desktop notifications. Awesome.
Wonder if I can hook that up to the rest of my desktop notifications via the
D-Bus NotificationItem-1 spec. That would be excellent.
Aww, crap. I wake up and there's 23 spam comments in the queue. They don't
even conform to my markup language. The websites they link to also don't
exist. It's like they're random robotic web vandals. Looks like it's time to
implement some anti-bot tactics. UPDATE: we'll see if this works.
My laptop's backlight is fading on the left side. This makes me sad, since I
have a white-on-black terminal that I usually keep on the left that is getting
increasingly hard to read.
This was actually pointed out to me at Mystery Hunt, but I think it's gotten worse.
Or maybe I'm just using my laptop in too well-lit areas. Or maybe my eyes are
worsening, and it's affecting me more. Either way, blah. I may try disassembling
the panel and seeing if it's just a loose wire or something.
In other news, Kubuntu fails at packaging KDE 4.6 - despite my best efforts,
the gstreamer Phonon backend (now made default!) and Pulseaudio are not getting
along at all. As in, I can crash both systemsettings, plasma-desktop, and
pulseaudio itself, all by trying to play a test sound. Good job, guys.
I shouldn't complain; rather, I should track down the bugs and fix them. I'm
happy to fix upstream bugs, but a tad frustrated with distribution bugs. I'm
considering installing Fedora on my desktop, since they manage to do a much
better job of shipping a working, fully-featured KDE. One prerequisite for
switching is verifying that I can (without too much trouble) do all the same
things that I used to do with Ubuntu. So I made a list, and am working through
it. Tasks remaining include running a DHCP server and verifying support for
GRUB2.
Last night, Jono and friends held a get-together to celebrate
Australia Day
(which actually takes place on the 26th, but this was the opportune weekend day). We
grilled sausages, watched
The Castle, and then
sat around and talked about all sorts of things. One of Jono's Australian
friends and I talked about politics and political theory for quite some time.
We covered health care, parents' rights over children, age of independence, and
the greater standard deviation of wealth in America as compared to other
prosperous nations (and in turn, its positive and negative impacts). We also
spent a while analyzing how the US and China are interacting financially, and
how we're both parasitic toward the other. I spent a good portion of today reading
up on the theories of John Rawls and Robert Nozick.
I'd forgotten how long it takes to copy 351 GB over the network. Even with
gigabit, cruising at 20MB/sec (which is rather optimistic, given that the other
end has to flush everything to disk as it goes, and there are plenty of small
files that will wreck performance), you can expect to spend 5 hours. Using a
10/100 switch will take about twice as long. I'm pretty sure it's still faster
than doing the same thing with a USB disk, though.
So, hopefully my backup will finish by noon (it didn't, but it did finish
within 15 minutes of noon), after which I can safely wipe my desktop, go crazy
with repartitioning, RAID, encrypted volumes, and LVM, and install Fedora.
Woohoo.
PG&E's automated caller called me at 8:00 this morning to remind me about my
appointment tomorrow. Yes, I haven't forgotten. Yes, it woke me up, and I
can't even be productively grumpy toward the computer.
Discovery of the day: on Fedora, simply setting "NM_CONTROLLED=no" in
/etc/sysconfig/network-scripts/ifcfg-ethX does not stop NetworkManager from
managing the interface. You also need to specify the appropriate MAC address
as the HWADDR. This took me way too long to figure out.
It also took me a while longer than I'd anticipated to discover that
NetworkManager 0.8 does not support ethernet bridges. While this was fairly
expected, I was still a little frustrated that I couldn't find this in the
documentation for a long time.
BiD lunch today was pretty awesome - not only did Anuj and Nick walk to
Cheeseboard and back to get six delicious pizzas,
Jonathan Bachrach came and gave a slick
talk on his ideas for the future of materials, CAD/CAM, and construction. He
presented a technique for constructing many types of solids by chained
tetrahedra by proving their satisfiability, doubling the voxel resolution,
incrementally merging paths into a single loop, and then exploding the solid
into a chain, tracking angles at each tetrahedron's bounds to produce the folds
necessary to fold the desired solid from a straight chain. It was like doing
protein folding backwards - start with what shape you want, know that it's
possible to make, and then work backwards to a starting point. The best part
is that each tetrahedron in the chain can be represented as a single bit.
Sweetness. Sample application: reconfigurable materials. Turn your desk into
a chair, a fan, an addition to your house, anything. I was impressed. One of
the coolest things was that his "slides" were not done in Powerpoint, or Keynote,
or anything that you normally do slides in. He had a full hour's talk prepared
as a long sequence of interactive 3d models, which he'd flip through like a slide
deck (but so much slicker). I am inspired to give awesome presentations.
I meant to go swing dancing tonight, but my garage gate opener is defunct (or
perhaps the gate motor or receiver is, or something), so I would have wound up
missing much of the lesson and not being able to give a ride to my compatriots,
so I decided to bag dancing tonight in favor of calling my parents, which
worked out nicely. I hadn't talked to them in rather a long time, and it was great
to catch up.
I think I've gotten everything that I used to run on my desktop functioning
again. Whether or not this will persist across reboots is something that will
be determined after I get home - I've learned not to fiddle with network
configuration and reboots remotely.
Oh, and I need to fix Fedora's GRUB2 numeric sort. Somehow, 2.6.35.10 is
sorted as "less than" 2.6.35.9, which is just flat wrong.
UPDATE: definitely did not configure things to persist across reboot. The
following sequence of events looked something like this:
Drove down to Palo Alto to interview with imo. Impressions: they are all REALLY sharp folks.
I felt I did acceptably on the first two interviews, although I needed a little
more prompting than I had hoped, particularly for the second problem. My
implementations eventually reached adequacy, though. For obvious reasons, I
can't discuss the problems themselves.
Had lunch with Robert Burke and John Rizzo (who both work for imo). We
went to this Chinese place and I had sweet walnut prawns, which were absolutely
amazing. I need to find more Chinese places that serve this dish.
My third interview (which was after lunch) was a bit worse - probably part
food coma, part harder problem. I got way more prompting than I should have,
although I eventually produced an adequate implementation.
Chatted with Robert for a bit on technology, life, the universe, and
everything. Also played his Tetris Attack clone.
One of these days I'll figure out how to make longer chains, and everything about that game will
just click, and it'll be amazing. So far, however, this has not happened.
Drove home. Traffic was very congested.
Computered (I'm making up words now) for about 20 minutes
Laid down for a "20 minute nap"
Woke up 4.5 hours later when Andrew called me, asking if I was going to Justine's get-together. Oops.
Went to Justine's get-together. There were eight people playing Bananagrams; I managed to win a round.
Sat around on the couches, discussed Geocities, Pokémon (and Pokémon
cards!), world records, the difference between NASCAR and Formula 1 racing
(including cubane), and the
.edu TLD. Good times were had by all. Quote of the evening: Greg, to Justine:
"That was way too high-pitched for us to understand what you were saying, but we got the gist."
Maybe it was a had-to-be-there moment.
I keep saying that neither of the dominant powers in the government are doing
enough to get our spending under control. At the risk of sounding pretentious,
I'm right.
So is Rand Paul.
Although as I've noted before, I'm a sucker for education and research, so I flatly
disagree with many of the cuts he's proposing, I think this is the first legitimate
proposal for spending cuts that comes close to the order of magnitude that we need to
get our budget under control.
I spent several hours today on Python Challenge.
It's kinda like notpron, except meant to be
solved with the assistance of Python programming. I learned more about the httplib module,
the regex module, the zipfile module, and the pickle module, so I'd say it was a decent
investment of my time.
Patent reform: you're doing it wrong (pdf).
Giving patents to the first inventor to file (rather than first to invent) means that large multinational companies with lots of resources
can take patented ideas from other countries and patent them in the USA. Smaller damages means it won't cost as much for the
big companies when they're found to infringe on each other's patents, but also means that it won't cost them as much when
they're found to infringe on small companies' patents. And what happened to prior art? The whole thing seems to make
patents worse at their original intent - to promote innovation by those who fear others would copy their ideas before they
can monetize them. This proposal weakens patents as the leveler between the small innovator and the large corporation.
This also sucks for academia, where you're supposed to publish as soon as
possible, once you have good results. Now, if you want ownership of your idea,
you have to get your patent filing in first, too, in addition to getting your
system working and reduced to practice. This could have the unwanted side
effect of forcing innovators to pick between publishing and patenting.
If you've been following the George Hotz vs. SCEA lawsuit (summary: Hotz
succesfully reverse-engineered and published an unrevokable key used in the PS3
security system, SCEA got mad and filed a lawsuit against him in California (he
lives in NJ) to make an example of him), **** just got
real. I think I'm
pretty well-informed on the facts in the case, and while I don't know all the
relevant US court cases, it seems pretty clear that Hotz hasn't broken any law.
This evening, Luke and Jake invited me over for some Settlers of Catan. At 10:15 pm.
'Twas awesome. After much distraction, we played a fine game in which I won by a single
turn at the very end, thanks to a Year of Plenty that I used to build a road and a settlement
on a desert space, which I followed up with a Victory point. It was a thrilling game.
After the round, we sat around and chatted for a while, and Jake played songs on guitar,
and then we played a few matches of N64 Super Smash Brothers. Jake is ridiculously good
as Link. Eventually, it was quite late, so I grabbed my tools (I'd left the set with Luke
from when we built our multitouch table early last semester) and headed home.
As part of my efforts to harmonize my home directories (and ensure I have an
offsite backup of important schoolwork), I've set up ownCloud on my dedi, which
I can mount via webdav with davfs2. It takes a good long while to upload 5GB
of data to my server via a residential modem.
I've also discovered that rsync with its default options doesn't play nicely
with davfs2 and ownCloud - not sure why, but TONS of files wind up in the
lost+found folder. Throw in the --inplace option, though, and everything works just dandy.
It's the strangest thing.
Also, based on my apache logs: it would probably be wise if I pushed the ownCloud
logs off into another logfile, 'cause MAN, WebDAV is chatty. I suppose there's not
really a better way to do file locking over a stateless protocol like HTTP, but
still. In addition, it'd be great if WebDAV supported ctimes and mtimes, because
something in the chain (davfs2/owncloud/ext4) is losing them, which means rsync
wants to overwrite every file, including the ones that have already been transferred.
This is fail, and merits further investigation.
Today was also Super Ad Sunday, a day during which (in accordance with
tradition) Americans watch hilarious ads for cars and other products,
punctuated by snippets of American football. Andy, the BiD lab grandpa,
invited the whole lab over for the afternoon, and there was talk of meeting to
go rock-climbing beforehand. It turned out that I was supposed to meet up at
Andy's apartment, rather than at the rock-climbing gym, so I was the only one
who went rock climbing, but I had a great time and thoroughly wore out my arms.
The gym does mostly bouldering, but also has a couple of slacklines, rings, and
a ladder consisting of doorframe-size holds. I learned proper climbing technique
from a couple guys who were doing ridiculous routes. I discovered that my body
is pitifully weak. I must resolve this.
As for the party: there was much food and drink, and about a dozen folks
gathered to celebrate. After the game was over, we played some Rock Band, and
I finally got to try out the keytar. I think I'd honestly do better with a
staff and sheet music. Maybe I just need to get used to the key/color layout,
and then it'll all make sense. Thinking about what key the songs were in and
the chord progressions they used made everything a lot easier. I could play
Pro mode on medium/hard, which I consider not bad for a first try. After
playing a bit on keytar, Lora made fresh ramen and I boiled gyoza, and then I
took a turn on the mic, singing "Saturday Night's All Right for Fighting" by
Elton John. Good times.
I've isolated the datestamp bug to davfs2 by testing the same rsync on a dav
folder served up by apache's mod_dav_fs, rather than by ownclould. Since the
result (incorrect timestamps) was the same, I've concluded that owncloud is not
necessarily the problem. Time to dig into the davfs sources and figure out what
the deal is with timestamps.
A couple random quotes:
"Take the recycling to the curb; a Jawa will take it."
"We should probably read the paper. Because the paper is peer-reviewed, unlike the lecture notes."
I sent my friend Sarah a surprise yesterday - a copy of The Upside of Irrationality.
Looks like it arrived. :)
According to John Canny, one of the annotated data sources used for speech
recognition is...reality TV shows with transcripts. My mind is blown - I
thought that reality TV couldn't be useful for anything. Apparently, I thought
wrong.
Now I want to think about other ways that something I consider useless could,
in fact, be turned to work for good.
Last night, I dove back into my broken high-resolution patchset for
libfreenect, the open-source Kinect drivers. This debugging trip took me
through compiling libusb from source to enable debug messages, finding that
libusb was getting a status code of -2, searching through the Linux kernel for
the matching ioctl(), and eventually finding that the command in question was
returning -ENOENT, which means
that the device completely disappeared. As a result, subsequent libusb calls
to the same open device handle would cause libusb to crash. Ordinarily, you'd
stop sending commands to the device as soon as it disappears, but the cleanup
code in our demo wanted to shut the device down cleanly.
It took a while for me to realize what this meant - through a particular series
of control commands to the Kinect, I was CRASHING THE KINECT, and it stopped
responding to further communications. After a couple of random tests, I hit on
the solution: don't send the "stop the video" command more times than you send
the "start the video" command.
Then there was another bug in my own code that had evaded me for two months
where, when starting a video stream, I'd queue up 16 isochronous transfers
before aborting because the pixelformat/resolution combination was invalid.
Obviously, I should have done the check up front and failed early, before
calling any functions with side-effects. If I'd followed marcan's software
development advice, I would
have figured this out a long time ago. He writes:
"One great technique to use is to do as much as possible in advance. Gather all
required information about the system, read any required data files, prepare
any modifications, and only at the very end actually commit the changes to the
device. If anything goes wrong during the preparations, you can just abort the
entire operation and be certain that the device is still safe."
Bingo. And as soon as I did that to my code, everything started working
consistently. Thus, I'm now in cleanup mode for this patchset, which I hope to
submit upstream some time in the next couple days. Many thanks to Antonio
Ospite for testing and reviewing my patches.
So, all that was last night. As for today: the noteworthy event was the
miracle berry social hour.
The CS department regularly hosts social hours, which usually consist of a
person or two getting a bunch of somewhat-themed foods and beverages, and then
bunches of people would gather and chat for a good while. This week, Justin
(who is one of my teammates for my security project) organized a social hour
themed around the miracle berry. The interesting thing about this particular
berry is that it contains miraculin,
a glycoprotein that fools your taste buds into thinking sour things taste
sweet. It's pretty crazy.
So, Justin had procured $60 worth of miraculin tablets and $40 worth of other
foods, like lemons, limes, oranges, tomatos, pickles, radishes, and so forth,
and they were all sitting neatly arranged on a table outside Wozniak lounge. I
let half a tab of the miraculin-cornstarch mix dissolve on my tongue, and then
ate about a third of an orange slice and nibbled on a lemon wedge. It was
pretty weird, having the lemon taste so sweet.
Then I started getting a headache, so I went inside the Woz to sit down.
Shortly thereafter, I realized this was no normal headache, but rather a
low-blood-pressure headache. I took off my backpack and put my head between
my legs, in an attempt to stay conscious.
Then the nausea hit. I hastily moved to a chair next to the giant trash cans.
Justin watched over me as I described as best I could that I was suddently
unwell, my blood pressure was dropping, I might pass out, no, I'm not diabetic,
no, I don't have any known food allergies, and no, I don't need an ambulance
quite yet, but if I do pass out, that might be a good next step. Then I puked
what little I had eaten up, which was a most unusual experience. Thanks to the
miraculin, the bile didn't really taste bad at all - it was mostly sweet, with
a very small hint of puke flavor.
By this time, I'd attracted a small crowd of folk who watched over me as the
color slowly returned to my sickly-white face. Justin brought me my backpack
and my "Powered by Open Source" hoodie which lay across the room. Another guy
brought me a cup of lukewarm water. I shivered in shock for a bit, but
remained conscious, and after a while, my fingernails turned from blue back to
their normal pink, and I was able to stand up and walk around again, after
which I went home and took a nap.
Other thoughts: I've had this particular response to stimuli three times in the
past six years or so. All three were after receiving flu shots, so I've come
to associate the headache, optional passing out, and optional nausea with that
particular immunization (although really, it's just vasovagal response). As
for what happened today: I'm unsure, but if I had to guess, I'd say maybe I'm
one of the rare people with a miracle berry allergy - the internet seems to
suggest that in rare cases, people get nausea and cold sweats after consuming
miraculin. I was quite proud that I did not pass out this time, and in
retrospect, this was perhaps one of the best circumstances I could
have had such a reaction in. After all, I was surrounded by intelligent
friends who looked out for me and entertained me while I recovered.
I registered this domain less than a year ago. Before then, I was practically
impossible to find on the Internet - I had a small website that I hosted as a
subdomain under my friend Matt's domain. It sat on a VPS that later went
belly-up when the admin managed to (in one fell swoop) ruin every client's
filesystem. Good thing we had backups.
Fast forward to today. In the past year (as far as the web is concerned), I've
created my blog and written up a couple of projects. I've made a brief landing
page. I still haven't gotten around to doing any design for the site.
Nonetheless: search for my name on Google
today, and I'm the top result (at
least, for all the boxes I have access to).
Lora invited me to join her Buffy the Vampire Slayer-watching group (which
consists mostly of friends from when they were all at Stanford). It turns out
they were in the middle of Season 5, which was not far from where I was in the
series. I agreed, and since I have a car and she doesn't, I volunteered to
drive.
The day held much more than just watching several episodes of one of the better
TV series of the late nineties and early noughts. Here's a summary of the day's
events:
Woke up after 5 hours of sleep, since I stayed up too late hacking
together a new viewer application for libfreenect. Ate breakfast (two english
muffins with cream cheese and jam). Picked up Lora.
Drove to the "secret Safeway" in Palo Alto to get food for the day. It's
pretty aptly named - I definitely didn't see the Safeway until we were already
in the parking lot. We got a variety of foods and snacks, including mango
sausages, carrots and hummus, and dutch crunch bread. Since asparagus is in
season and on sale, we got two bundles for fantastically cheap. Oh, the drive was
lovely - there was fog, and then right after we passed through a tunnel, the scenery turned
bright and sunny and warm and sparkling and wonderful.
Went to an apartment in Palo Alto, where I met Lora's friends Bradford (the
apartment resident), Kat (who was visiting from UT Austin), and Ben. We had a
delicious lunch, with the sausages, dutch crunch bread, roasted asparagus with
minced garlic, and carrots and grapes. Our viewing took us through roughly the
middle third of Season 5 of Buffy. We noted that the TV made all the motions
look incredibly smooth, which was strangely unsettling. Maybe I just like
encoding artifacts, or low contrast, or something else ridiculous. Joss Whedon is
masterful with the use of silence as a sound effect. Around
4:30pm, Kat had to leave to pack her stuff and catch her flight, so we said our
goodbyes.
Drove to Lora's parents' house, where we searched the attic for (and
successfully extracted) an electric keyboard. Backstory: the BiD lab is
putting together a rock band, and we needed to bring in an electronic keyboard,
since both Lora and I play piano (although we're both only classically
trained). Lora owns such a keyboard, but lacking a vehicle, had no easy way to
bring it to campus. We found the keyboard and stand, although we couldn't
find the power supply. Lora exchanged seeds with her parents (they all keep quite lovely gardens), and we picked a
bunch of oranges and clementines from the trees in the backyard. I plan to use
some of the oranges for juice. We also worked out that Lora's iPhone is on the
fritz, but that her SIM card is not, since the SIM worked in my N900. Oh yeah,
Lora had been calling her friends by looking the number up on her phone and
calling from mine all day. We gathered up all the stuff we had collected at
Lora's parents' house, and went...
Not home (yet). We called up a couple more of Lora's Stanford friends, and
met up at Soja's flat in Mountain View. I recognized the complex that
Sukrit lived in last summer as we drove by,
and also noted the tasty Thai restaurant I'd visited with the MIT folk.
Soja is helping write a Bay Area puzzle hunt (in the Stanford style, not
the MIT style, but hey), which I got excited about. I love puzzle hunts!
We ate leftover quiche and hard cheese with quince jelly, and sat around
and chatted with the 8 or so people (most of whose names I've forgotten
>_<) until right about 8pm. Then, we...
Drove back to campus to drop off the keyboard. My mom called me on Skype
to wish me a happy Valentine's Day while I was driving, so Lora answered my
phone and put it on speakerphone so we could all chat, and we talked briefly
over the 3G. We bet on who we'd find in the lab at 9pm on a Sunday, and we
both lost. Lab grandpa Andy and Kim were there, and Andy had lined up four of
the tables together and set a net across the middle, producing a rather nice
improvised ping-pong table surface. This was an EXCELLENT use of the lab
tables. I played a match against Andy and narrowly lost while Lora printed off
a copy of whatever notes she needed for Monday. We found an
appropriate-voltage power supply somewhere in the BiD filing drawers/cabinet,
and we verified that the keyboard was in working order.
Finally, I dropped Lora off at her house with the rest of her stuff, and went home.
It was quite the fun several-destination trip, and I met tons of friendly people. Yay!
My laptop's backlight went out again today, and stayed out. This was quite
unfortunate, as I was expecting my laptop to last me a good while longer. So,
since I had planned to spend my Monday with a functional laptop (since the
Qualcomm Innovation Fellowship app is due tomorrow ahhhhhh!), I spent a couple
hours with the best screwdriver I've ever
used,
and tore my laptop apart. Wouldn't you have done the same? :D
A couple pics:
Laptop in pieces on the table.
Backside of the monitor, once removed from the chassis.
Close-up view of the high-voltage backlight power connector that I disconnected and firmly reseated.
I put it back together, and boom! Backlight was back, and I could use my
laptop once more. It wasn't how I had planned to spend my Monday, but it
certainly beats being out a computer.
My high-resolution support patches have landed in the unstable branch of
libfreenect, which means I've broken everything thoroughly. On the upside,
I've gotten some great feedback
on my patchset, including at least one successful tester on
Windows (which I never so much as tried to compile on), and some great
discussion on acceptable API changes.
I also learned today that the guy who has taken a leading role on maintaining
the wiki and FAQ for OpenKinect as well as answering lots of questions on IRC
doesn't even have a Kinect. Nonetheless, he manages to help bunches of people
with their issues. He's just really observant (he's pointed out
inconsistencies in our documentation that I missed) and remembers what other
people say, so he can share solutions to the same issues. I am thoroughly
impressed. bankP (his IRC name) is a wonderful example of the fact that you
don't have to be a software developer to contribute to and make an impact on
the open-source community.
Therehavebeentonsofarticles
on the recent series of events regarding Anonymous and HBGary. While I won't
go into the details of how the security firm got their servers hacked and tens
of thousands of emails leaked (by a group of 5 people, including a 16-year-old
girl), I do find it worth discussing what has come to light as a result:
corruption.
I'm alarmed that our government and corporations are stooping to such tactics.
I've got so many questions. What has happened to the concept of honor, or
integrity? Where is transparency and accountability? Why did it take the
efforts of a group of hackers engaging in retaliation to stumble upon this? Is
this an isolated instance, or are these sorts of tactics systemic in our
government?
Finally, one more question I'm almost afraid to know the answer to: if we had
more transparency in our governments and corporations, what else would we find?
My friend Nick Zentay got into grad school, fully funded, at Texas A&M! I
am proud of him.
Taint tracking has to be the most underwhelming security tool I have ever heard
of. It simultaneously produces too many false negatives and too many fakse
positives. Any attept to reduce one increases the other, and both mistakes can
be easily triggered by benign code and malware alike. It balloons completely
out-of-control, unless you designed the whole program you're analyzing with
taint tracking in mind, in which case you already reap the benefits of
privilege separation thanks to a clean architecture. I'm a little frustrated
that we spent so long in Security today talking about such a fundamentally
flawed idea. Time to read some refutation papers (and then the refutations of the refutations).
Last night, I watched IBM's Watson defeat the two winningest human Jeopardy
players in the show's existence. It was one of the most impressive things I've
seen in recent history. Best part: during final Jeopardy, Ken Jennings
writes his answer:
Unrelated: my laptop's backlight now provides even lighting. This was not the
case on January 29th, when
I wrote about how
the left side was fading, nor was it the case when I was in Cambridge for the
Mystery Hunt, when someone pointed out that my monitor was kinda dark on the
left side. But it appears to be fixed now!
I wrote a not-even-half-baked log-parsing script so I could go through my
website logs and see how people were finding my site. With all due
respect, I'm not interested in how my friends arrived at my site - they
generally either have the URL bookmarked or in a feed reader or some
combination thereof. No, this is targeted at the people who don't know me
and wind up on my site (and hopefully find it interesting). I analyzed
only the most recent week's logs: Feb 13 - Feb 19. I have my site's
lifetime of logs, though, so I could feasibly run this analysis on the
entire history, if I automated it a tad more.
I get the vast majority of my non-self-referred traffic from Google. This
does not particularly surprise me. What does surprise me, though is some
of the search queries that manage to land on my site. So I went back to
Google and performed the same searches, and noted which result my site was,
at the time of writing. All query strings are reproduced verbatim.
Google query string
Google ranking
Referred page
Comments
18F2550 linux
7th
http://zarvox.org/usbsnes/
Whoa. A chip name and "linux" gets you my website? Crazy.
Based on HID-Test by Christian Starkjohann
4th
http://zarvox.org/usbsnes/
So, that's the guy who wrote the V-USB library I used for that project. Does it merit a top-10 result? I doubt it.
drew fisher
1st
http://zarvox.org/
We already discussed this one :)
cr48 cut and paste shell to chrome qemacs
4th
http://zarvox.org/shortlog/month
pollingg routine in assembly language
4th
http://zarvox.org/usbsnes/
Spelling mistake is faithfully reproduced
nick peach hbgary
22nd
http://zarvox.org/shortlog/month
This is what happens when a bunch of otherwise unrelated blog posts get indexed together.
People seem willing to go through a lot more images than regular web results.
pg&e fee emergency power hook-ip
???
http://zarvox.org/shortlog/month
I have no idea. I couldn't find this result.
atmega328p
Images: 6th
http://zarvox.org/ddrpad/RedArrow.png
For some reason, I get a surprising number of Google Image results for the red and blue arrow schematics from that project.
Use Xbox 360 Dance pad on stepmania
7th
http://zarvox.org/ddrpad/
Here's what you do: you plug it in. That's all.
atmega328 led internal resistor pull-up
8th
http://zarvox.org/ddrpad/
Drew Fisher
1st
http://zarvox.org/
Note the capitalization.
usb hid controller parts
9th
http://zarvox.org/usbsnes/
drew fisher
2nd
http://zarvox.org/resume.pdf
This one found my resume, which is a good second result after my website for my name, methinks.
v-usb hid report descriptor size calculate
34th
http://zarvox.org/usbsnes/
pic micro snes controller usb
5th
http://zarvox.org/usbsnes/
And I didn't even use a PIC for that project.
how can i know if my usb snes adapter works
2nd
http://zarvox.org/usbsnes/
Try it out and see? Also, really? 2nd?
miraculin and effecting taste
????
http://zarvox.org/shortlog/month
Pedantic: wrong use of "effecting"
light dance pad
10th
http://zarvox.org/ddrpad/
I happen to know I'm #1 for "LED dance pad"
rock band keytar python script
????
http://zarvox.org/shortlog/month
I couldn't find this one in the results. Color me surprised.
at90USB breadboard
37th
http://zarvox.org/usbsnes/
I guess I mentioned that chip on my page, and how I didn't use it.
Drew Fisher berkeley
4th
http://zarvox.org/resume.pdf
This search came from a Mac user in Germany. The first three results were from a class wiki and the BiD webpage.
I also had one user manage to arrive at my site via Yahoo, with the query "arrow LED using microcontroller" which found http://zarvox.org/ddrpad/.
I had one visitor from a Windows 98 machine.
Today I learned the proper name for the condition where you have eyes of
different colors: Heterochromia iridum. One of my friends
in middle school had one golden eye and one brown one.
The battery in my UPS is failing. As a result, it makes this obnoxious beep at
regular intervals. I've decommissioned it in favor of my (epic) 12-port power
strip, at least until I can get a new battery or new UPS.
Hehehe, Johnny Lee
(of Microsoft Kinect fame) was behind the Open Kinect bounty.
Apparently he wanted PC support for it when he was working on the project at
Microsoft, and management didn't. So he hatched a plan to get that technology
into the hands of the people that might create awesome stuff with it, and with
the help of the folks at Adafruit and a
worldwide community of hackers, he showed the world just how
cool NUI technology could be. Microsoft caught on, and has now
announced a development kit for researchers and enthusiasts for Kinect
on the PC this spring.
I'm a little creeped out. I would guess that either 1) someone inside BofA is
looking for info about their contact with HBGary, or 2) someone inside BofA is
looking to see what the Internet currently says about them.
In other news, Kenghao and I made some progress on our research. I
implemented a
root-finding algorithm
in C, and optimized it over a moving window of functions by applying
Newton's method.
Initial benchmarks showed my approach to be about half as fast on my laptop as
just solving for all roots explicitly with the BLAS eigensolver. On Kenghao's
machine, it was even worse - my code was slower by a factor of four.
Then we tried it on a Nexus One, and my code was 37% faster than the
eigensolver. Since this is intended for mobile phones, eventually, this is
promising. Later, I remembered that I hadn't enabled compiler optimizations,
and even on my computer, performance shot to within 20% of that of BLAS. I
still have some ideas on how to make it even faster.
Gah, algorithms problem sets SUCK. On the upside, I'm getting lots of good
practice writing things in LaTeX - real LaTeX for my homework solutions, and
mostly real LaTeX for equations in Google Docs, where I take my notes (gotta
put that CR-48 to good use!).
Reza brought his Vizsla (named Keplar!) into lab today. He loved playing
fetch. I could probably get along well with a Vizsla, or at least this one -
he was very quiet and unimposing, yet friendly and energetic. Usually I find
dogs either too loud and whiny or too slobbery, but Keplar was pretty chill.
Reza and Kenghao and I went through a bunch of piles of old electronics from
research projects past. Among various other things, we found several USB
soundcards with 5.1 out, optical out, and optical in, and they work
out-of-the(-carboard-storage)-box under Linux. I couldn't help but think of
Gabe and his awesome mixing with his falling-apart Mac and that unusually-nice
external sound card. Maybe if I have free time again, I'll play with
mixxx some more - the ability to DJ properly
wouldn't be a bad skill to have.
I am not needed for jury duty at 8am after all. I will check back between 11
and 12 to see if I am required for roll call at 1pm.
UPDATE: wasn't needed at 1pm either, which is an excellent thing, since I needed
the sleep. Still super tired.
TODO: write letter of recommendation for A&M prof, reformat lecture notes
for security class (before lecture), reimplement my Newton's method algorithm
with additional intermediate steps to make it work more of the time, and
spend some quality time looking at qemu's internals for my security class.
Politics linkdump time:
Supreme Court says (pdf)
AT&T and other corporations don't have "personal privacy" rights the way
individual citizens do. I agree - without complete knowledge, the free market will not
avoid immoral behavior, even if that be the desire of the citizens. I could take
that and turn it into a full-blown discussion, but I don't think I have the time nor the
coherence to write that up right now.
Republicans still don't get net neutrality.
The point is to prevent discrimination based on content - preserving that
freedom of speech that you keep hammering on and on about. Adjust prices based
on the quantity of bandwidth used, if you like - that's perfectly reasonable
usage-based billing - just make sure you market it like that. I could do a
proper rant for this too, but again: sleepy, and too much TODO.
Wow, I'm really behind. I've only written once in the past week. Ouch.
Wednesday: Leila Takayama
of Willow Garage is open-sourced for the benefit of all. Their goal is to make
personal robotics a reality. They've also been really excited about the Kinect,
and are apparently interested in getting an open-source driver for the microphone
array. Say, that's on my TODO list!
As any marketing person should be able to tell you, "under-promise, and
over-deliver" is the correct approach to robotics. Tell people that the robots
are really dumb, and they'll think much more highly (and significantly so) of
the same robots than if you prime them with the belief that they're smart.
Also: robots (currently) take a long time to think when processing huge amounts
of sensor data. During this time, they appear completely uncommunicative.
Just adding the gesture of "scratching" its "head" while doing this computation
made test subjects think more highly of the robots and their intelligence.
Fascinating. To make robots more human, we need to give them idle movement.
There was also a good bit on the Texai, a set of telepresence robots. If you
watch the Big Bang Theory, you might have seen
this episode
in which Sheldon uses a Texai to represent himself.
Sadly, since I have class at 1, I had to leave just before the end of Leila's presentation,
but suffice it to say: it was pretty cool stuff.
Then in security class, we had a guest lecture from Google's
Adam Barth, who is basically the leading
mind in security on the Web. We discussed how the Web originally had no security policy,
the development of the Same-Origin Policy,
how cookies are broken, cross-protocol attacks, and what can be done in the future to
improve the security of the web. This was also excellent.
I spent the evening writing a letter of recommendation for one of my professors
back at A&M who was nominated for a teaching award. I thought back to
writing my statement of purpose for graduate school: evidence-based anectdotes
that show what you're saying are more convincing than simply saying something.
By the end, I thought it was a pretty good letter. Good luck, Dr. Liu! Thank
you for all the guidance and preparation you've given me.
Thursday: I woke up at an unholy hour of the morning to join my fellow BiDizens Wes, Andy, and Lora
at Bridge's Rock Gym in El Cerrito. I had MUCH more
success at climbing than the last time I was at Bridge's. Perhaps it was the moral support, or
perhaps my arms have grown stronger since, or perhaps it was because I ate a good breakfast.
In any event, I had a ton of fun, despite scraping skin off my fingers like a cheese grater.
I'll grow some calluses if I do this often enough, surely.
After climbing until my arms were doing that thing they do where I can no
longer grasp anything long enough to get both feet on the wall, I went home and
took a nap. It was excellent.
Algorithms class was uneventful, though I did get a couple more humorous quotes
from Satish Rao. For the record, this is the professor who's said:
"Somebody, say, God...or a linear programming package gave you this solution"
"Now we know: everybody makes mistakes." (related to the experts algorithm)
"So that's the way bigger works. Okay, good!"
On doing random walks to select vertices of a graph uniformly at random: Justine: "I'm lost again". Satish: "You're supposed to be, so that's good."
After losing a constant due to shifty algebra: "There's some way of retrieving the other factor of two, but don't worry about it." (This is actually okay; in algorithms, since you generally work with big-O notation, you don't care much about the constants. It drives engineers mad, though.)
"Random bits are a gift from God. And you can read them in constant time!"
Student: "Can you use a marker with higher contrast?" (Note: the marker currently in use is blue) Satish: *looks at marker* "Is there a contrast button on this thing? What other colors do I have?" (Note: there are 7 markers on the board, and 6 of them are blue) Student: "There's a red marker..." Satish: "Red is alarming!" *draws line with red marker* *jumps and gasps*
The other professor for that class, Umesh Vazirani, has also produced some
excellent lines:
"Why do we do this? This is just what theoreticians do!"
"So, back to this problem you should all be tired of by now..." (after we spend a fifth lecture discussing the congestion-minimization max-flow problem)
"Back to how soon we get lost."
"Theorems, unfortunately, don't work this way. You actually have to prove them."
"We are almost done! Except for about 40 pages of the paper..."
Umesh: (Long stream of strange calculations) "Then we can just use Cauchy-Schwarz!" *blank stares from class* Justine, on behalf of some 80% of the class: "What is that?"
Umesh, for the third time in a single lecture: "We're almost done!" *erases half of the giant whiteboard*
It may be confusing at times, and the problem sets may suck, but at least the
class doesn't lack for entertainment.
That evening, Andrew had organized a potluck dinner party at Luke's apartment,
so I made a giant bowl of cucumber salad with many colors of bell peppers and
grape tomatoes and mozzarella. Other folks brough roast chicken, chips and
guacamole, pita bread with feta and hummus, collard greens, and Luke made this
excellent chicken noodle bake topped with melted cheese...all was excellent.
It turned out that the garage door at my apartment complex was defunct (again!)
so I couldn't drive, but fortunately David and Shaddi were already heading our
way to pick up Jon, so we all packed into the '95 stick-shift and successfully
traveled to our destination. (Fun fact: that word can be spelled "traveled"
(US) or "travelled" (UK) and both are considered correct.)
Then I got home, reverse engineered the Kinect audio firmware upload procedure,
successfully extracted from a USB dump a byte-for-byte copy of the firmware
originally found in an update package, and briefly wrote up my work and sent it
to the mailinglist before realizing that it was 3am and OH CRAP I NEED TO MEET
KURTIS AND BRIAN IN LAB IN 6 HOURS SLEEP TIME.
Friday: Overslept, slightly. Was still was the first one to lab - "Berkeley
time" means 10 minutes after when you said something was going to happen. I
worked with Kurtis and Brian on some electronics, verifying that various
components in the circuit were functional, and trying to hunt down the reason
why the vending machine motors were not turning at a satisfactory rate.
Hastily prepared an application and NDA form in advance of my pair of
back-to-back Google interviews, emailed the signed and scanned pages off to the
recruiter. Why they let you sign the docs with a self-signed Acrobat reader
cert but not a PGP key in the strong set is beyond me. Then I ran home, since
I have no cell reception in the BiD lab, and no good place to sit for a phone
interview on campus anyway. Interviews went seemingly well.
After the two hours of interview (during which time I was very very hungry) I
made myself some lunch (bagel and lox + leftover cucumber salad) and settled
down for a rather long nap, which resulted in me sleeping through a meeting for
my security team. >_>
Time to implement some code for research, then work out patches for libfreenect
this weekend.
So far, there has been little public work on developing drivers that work with
the Kinect NUI Audio. For the unfamiliar, the Kinect has a four-microphone
array that can be used to identify the direction from which a sound came.
Further, the Xbox360 also sends it whatever it's playing out the speakers in
real time, so the Kinect can cancel that out from the microphone data. It can
hear you (just you!) over the explosions or whatever sound your stereo is
playing. That's pretty cool.
However, none of this has been available outside of the Xbox360 SDK. I
think that some people would find this useful, and that even after Microsoft
releases their development kit for academia and hobbyists, some of these folks
might want to use open-source software.
So: here's NUIse, a
proof-of-concept driver for the NUI Audio device. I reverse-engineered the
protocol and implemented my own software that streams the data from the four
microphones to files. I'm in the process of documenting the entire protocol
on the OpenKinect wiki.
Since it's hard to post a screenshot of an audio driver, you'll have to settle for
this audio recording.
This was not accomplished in a vacuum. I owe thanks to a number of folks that
made this work possible. To name a few:
Adafruit industries (and their motivator, Johnny Lee) for recording the sole USB log (and video) that I used to reverse-engineer the communications protocol
Sebastian Ortiz (trtg) for writing up what he found and suggesting the 4-window 16kHz 32-bit PCM audio interpretation
Microsoft, for making this awesome hardware in the first place
Leila Takayama, for telling me that Willow Garage really wanted to see someone work this out
Sometimes I wonder if it's weird that I only really have one drawer of files.
Financial records on the same level as middle school papers.
Tonight, Andrew and I are hosting a dinner party. I'm supposed to make bread
bowls. I hope they turn out well.
I've also been thinking: are there tools to empower or assist manual reverse
engineers? I basically stared at hexadecimal digits until they made sense when
working on the Kinect, but are there tools to do the same sort of
pattern-matching that happens in my mind? Could they suggest structure and
interpretation of the data? Would they actually be useful, or only an academic
novelty?
I am in the process of writing up an explanation of how I worked out the Kinect
audio protocol. I intend for it to:
Show you the way I approached and solved the problem
Share some insights and techniques that might be reusable for other reverse-engineering tasks
Be at least marginally entertaining to those interested in the topic
Sebastian Ortiz posted more USB logs, including some during calibration, so I
may be able to get a better understanding of how the noise cancellation works.
That will make the aforementioned article longer. Oh dear.
UPDATE: bread bowls turned out AMAZING. Also, I need to learn to get over the dough
being sticky on my hands and add more water to bread dough in general.
Today I discovered Tax The Churches.
I'm curious as to what other people think about this idea, but also about some
of the background behind taxes and tax-exempt organizations. I had some other
thoughts, but I couldn't express them clearly, so I opted not to write them.
Sometimes there are things that I want to say, or write about, but I'm afraid
to make them public for fear of the social consequences of saying such things.
Thus, I'm planning to take after roc, write my thoughts anyway,
and publish only the hashes.
If, someday, I reveal the posts, you'll be able to verify that I had written
them at the earlier time. I doubt these hashed posts will be common; I've got too
much to do anyway. :P
Yesterday, we started Daylight Savings Time. Since everything there still
applies, please read my objections from last fall.
CS Visit Day has been going well; more on that later.
Congratulations to the TAMU CCDC team for winning ANOTHER regionals competition.
You guys rock!
I find the Programmer Competency
Matrix
to be excellent. I'm proud to say I fit in Level 3 of a majority of the
categories. The nice thing about that visualization is that it reminds me that
I still have much to do and learn, and that I must continue raising the bar.
CS Visit weekend: a rousing good time. One of my hostees is definitely coming
to Berkeley, and I think there's a decent chance at the other. The potential
additions to BiD are also fairly cool people.
There are a bunch of BiD folk going to CHI 2011. We
may or may not be getting a penthouse for the week.
Mark Browning sent me a link to this video about
CCDC that I was (briefly)
interviewed for. I had completely forgotten about it. Mark and I appear together at the 28 second mark.
I heard and gave feedback on Anuj's practice quals talk today. His proposal
relates to children learning language through asking questions. I need to read
more developmental psychology. Mom, friends, moms of friends, and friends of
Mom: do you have any good books on that subject that you'd recommend?
Later, I went to a "Microsoft Tech Fest" where I wound up knowing more about
the Kinect than the MS rep. I was more than a little disappointed, but in
retrospect, I suppose I went into a recruiting event hoping it would be like a
tech presentation, which is my fault.
Accomplishment: I have reworked the libfreenect API and sent my patches upstream.
We'll see if I can figure out Cython so I can fix up the bindings. I also need
to figure out a way to import nuise into libfreenect. Augh.
On that note, though: nuise now detects if it's already uploaded firmware.
This means you no longer have to unplug and replug the Kinect (or change source
code) every time you want to run nuise. YAY FEATURES.
It rained a lot today. Hopefully my Vibrams will be dry by morning.
Interesting read of the day: a guy
posted his genetic code
on Github, a site for collaborative software development.
Someone else uploaded said code to Promethease,
which is a website that analyzes your DNA sequence and tells you interesting things about risk factors or ability.
Now I want my genome sequenced!
One problem with unified message notifications on my phone: when two different
people send me instant messages, they appear as the same notification window.
When I select it, it opens only the last message sent. This means that if
someone sent me an SMS, and then someone else sent me 7 lines of instant
message, it's very easy for me to not notice the SMS. This happened today.
Desired behavior: spawn a new window for each contact you receive messages
from. This will prevent messages from going unnoticed at the sender level.
Each window could receive the yellow "requesting attention" highlight. In
fact, this behavior would be nearly trivial to implement, since it's already
standardized and specified.
I wonder what the current behavior in MeeGo is. Might have to look at the
images again. UPDATE: no wonder my MeeGo experience was so bad in the past.
My SDHC card is only Class 2. They recommend at least Class 6. This would
vastly improve I/O (by a factor of at least 3), which in turn would improve
performance of any operation that is I/O bound, like most things involving
photos, or launching applications.
Oh man, awesome food discovery: Pineapple sausage
from Evergood Fine Foods. Absolutely delicious. I think I could eat these every day. And
they're sold at the Safeway just up the road from my apartment. WIN. By the
way, Mom, apparently the Costco by the Turnpike and DNT sells these; try them
out next time you're up there.
Random question of the day: do you think I am more sheltered or less
sheltered than you are? Why?
An hour or so this evening: ead2c3784f90a67f57de4baae27f3729. Although I do
need to stop fooling around and get back to finishing that post on
reverse-engineering the Kinect audio protocol.
Andy had some folks over for a St. Patrick's Day (observed) party. We totalled
8 - Andy and his girlfriend Kim, Kim's friend Judy, housemate Rachel, former
BiD student Seth, and I drove myself, Wes, and Kim from the BiD lab (same name,
different person). We had TONS of corned beef, cabbage, and carrots, with
Guinness Extra Stout. I baked a tasty loaf of bread, and Wes made soda bread.
There was much food and drink and merriment.
I also had passed Games of Berkeley on my way to the FedEx shipping center this
morning (mailing back signed forms for my summer internship), and decided I
would find Pandemic, a board
game that I played with Nick and Eric over winter break and rather enjoyed
(though I neglected to mention it in the relevant blog entry).
So Andy, Wes, Judy, and I played Pandemic. I forgot the win condition,
though, and thought that although we had collected all four cures, we had to
eradicate all the diseases as well, so at the time, we kept playing until we
ran out of cards, seemingly clutching defeat from the jaws of victory. It
wasn't until after the fact that I realized we had actually won. I had fun,
either way, and want to play that game more often.
The group also played a bit of Dance Central with Andy's Xbox360 and Kinect.
It turns out I do know someone in town with both pieces! It also turns out
that Judy is delightfully nerdy, has been
following the OpenKinect project, and has been super excited about it! Suffice
it to say that her reaction was priceless when I shared that I'm one of the
devs on the project. :) It was really neat to randomly meet someone who was
familiar with and appreciated my work. I hope we get a chance to meet again -
she seemed like a pretty cool friend to have. As a bonus, Andy let me confirm
my suspicions about the Kinect audio calibration routine by letting me run it
and listen to his speakers, so I can say for certain that I have the audio
channels labelled correctly now. Who says you can't have a party and make
project progress at the same time! :D
I also wound up teaching Andy's Kim how to French braid hair - Kim has really
long, thin, straight blonde hair that braids very nicely. Judy has layered
hair and bangs that made it more difficult for me to demonstrate on, but her
braid turned out decently too.
Then we played through New Super Mario Bros Wii. I picked up the controller
somewhere in World 2, and we played through to the end of the game, which took
until shortly after 3am. Practically the entire time was filled with laughter,
which is an excellent way for a party to turn out. Fun activities, interesting
people, delicious food, and laughter. What a great way to start Spring Break!
UPDATE: I have now implemented email updates for comments. If you don't want
to receive notification of reply comments via email for posts you have
commented on, let me know, and I'll get around to implementing that feature.
Spring Break has been mostly uneventful. I've not travelled anywhere, but I
have spent some time reading, sleeping in, working on various brain activities,
and generally relaxing.
Today I discovered a leak in the ceiling of my kitchen. Joy. I've notified
the landlord, but there's nothing that can be done until it stops raining.
(It's been raining on a daily basis for approximately two weeks now.)
I've continued work on the Kinect audio protocol, and I now believe that I will
have to correctly implement the cryptographic exchange to enable the noise
cancellation. While I don't think I need a private key for this, since every
Kinect has different certificates, I will be unable to perform a replay of any
of the USB logs that I have available, since they were not collected from my
Kinect. I must figure out how the computation was performed, and mimic it
fully. This is troublesome, but hopefully still doable. I'm learning about
ASN.1 encoding and even more about authentication protocols.
I recently received a card in the mail from my friend
Caitlin. Thanks, Caitlin! It
really made my day.
I've also spent some time working at solving
The 2 Tone Game, a mini puzzle hunt.
There's a Stanford-style hunt on April 9th that I'll be doing, and I might also
try DASH3 on April 30th. We'll
see how busy/hosed I get.
It is one month until my birthday. What I shall do that Sunday, I do not know.
Suggestions?
After a good while of sitting on our hands, Matt, Jason, Travis, Kyle, and I
upgraded our server from Debian Lenny to Debian Squeeze. As a result of some
incompatibilities, my blog had about an hour or so of downtime while I rewrote
the engine against web.py 0.3. Everything should be up again now; let me know
if you seen anything broken.
I notice subtle changes in the Gmail interface. Today, they made the green and
red status icons used in the Chat contacts list less saturated. I probably
would have missed it, except for the fact that the Labs feature that shows
little robot icons for people using Android phones still shows in the original
colors:
This was interesting, the most exciting thing I've seen in a while was this:
18:36 <@qDot_> BTW, zarvox, would you like integration rights on the repo? :3
18:36 <@qDot_> BEcause me doing this is getting a little silly, obviously.
18:36 <@qDot_> There's really no need for me to be man in the middle at this point.
18:39 <+zarvox> qDot_: I would love commit rights to the repo :)
18:39 <@qDot_> Ok. I'll talk to Josh.
18:39 <+zarvox> Thanks.
(a half hour later...)
19:09 <qDot_> Congrats. You're a maintainer as soon as Josh gets you added, which will hopefully be in the next few minutes.
19:13 <zarvox> :D
19:13 <zarvox> *gives qDot_ a big hug*
19:14 <qDot_> Something we should've done a long time ago. :)
19:14 <zarvox> heh, I figured I had to show I was reasonably reliable and going to hang around and that my code didn't suck first
19:16 <qDot_> Oh, no, it's more than we're all just flakey and once you get in the routine you forget you can change it or that other people can help.
19:17 <qDot_> But, yes, what you said too.
So, now I'm official an OpenKinect maintainer. Sweet. Thanks, qDot, JoshB,
and marcan! You guys rock.
I'd thought for a while that I really ought to make a significant contribution
to the open-source world. I'm reaching the age that a lot of the folks who
wrote pieces of software that I admire were when they made their contributions,
and I'd been feeling like I should be doing the same. I've been programming
for over 10 years now. It's high time that I made something useful and gave it
to the world. So, one of my personal goals has been to produce a useful piece of
open-source software.
I'm on that track now. I'm not done. This is only the beginning. But I'm moving
in the right direction. :)
Sebastian sent me another USB dump, which I'll use over the next couple days
(read: weeks) to verify some useful things about the Kinect audio stream. I'll
probably either wind up implementing a decent portion of TLS 1.2 to get this
working, or I'll use gnutls or some other library. Trouble is, gnutls is
LGPL,
and I was hoping to relicense this code under the
2-clause BSD license.
I'll figure something out. Probably much later than planned,
though - this week is going to be busy. UPDATE: JUST KIDDING GNUTLS WILL BE FINE.
I fail at remembering the details of the LGPL.
I woke up for the first time this morning at some hour like 07:30, glanced at
my alarm clock, and rolled back over and went back to sleep. I woke up a
second time at 10:10 when my mom called me, asking after the whereabouts of my
lockpicks (one of her friends managed to lose the key to her lockbox). I woke
up for a third time at 11:50, when my phone kindly reminded me that I was
supposed to be on campus by 12:05 for lunch with prospective M.Eng. students.
That last one inspired fear of tardiness in me, and I rather quickly made my
way to campus. Of course, I arrived right on time. :)
I have a frightening amount of work to do this week. A few papers, and
implementing process separation for qemu I/O
devices for a class project. Time to get intimately familiar with a pile of
UNIX socket and syscall functions.
Also time to either implement a GUI in C++/Qt for exploring a dataset I have,
or figuring out the Python bindings to clapack. While we're at it: does anyone
know which is more draining on a cell phone battery: 100% CPU usage, or
constant 3G radio usage? This project has a design tradeoff that probably
depends on these numbers.
Also time to read some papers for security class.
Hmm, the word "time" appears eight times in this post (ten if you count the
word "times"). Guess what's on my mind lately?
Comcast is having an outage that includes my apartment. Annoying. I need to
check if they have any semblance of an SLA in their service contract; the
amount of time their service is down is starting to get ridiculous.
Good thing I have my N900 and my CR-48, each with service from a different
provider. Redundency is a good thing.
UPDATE: Oh, the irony. Right as I finish writing this, service comes back
up and nagios sends me an SMS to let me know.
Algorithms class moving substantially faster than usual, due to your favorite question-askers being out of town at a conference.
GNUTLS version whatever's-in-current-distros not supporting the ciphersuite you need, forcing you to compile the most recent version from source
GNUTLS version the-most-recent-code-written not supporting the certificate format you need, forcing you to either implement it in GNUTLS or hack something else together yourself
Things that are awesome:
In-N-Out. Three double-doubles. So tasty.
Writing your own bignum library from scratch and having it work
Catching up with friends you haven't talked to in too long.
I feel like I should have some examples of things that are more than awesome,
to demonstrate the axiom of trichotomy. Maybe
some ridiculous April Fool's day joke will qualify.
I baked cookies with white chocolate chips and walnuts. The recipe I used was
ridiculous, and called for too little flour and too little baking time. The
first batch turned out a bit messy, but the rest were all right (after I added
more flour to the dough and baked them for two more minutes).
For the record, the weather is downright gorgeous today. Cool and breezy, but
with a nice sunbeam. The serenity makes me want to just sit outside and let
my mind wander.
TODO:
File taxes
Algorithms midterm (ugh)
get process isolation working in QEMU
get NUIse to build/run inside libfreenect
release libfreenect 0.1
I may use libtomcrypt
to implement the cryptographic authentication involved for the Kinect (since
GNUTLS doesn't understand the cert I have to deal with), but that's getting put
off until later, on the grounds that I don't have the time to learn Yet Another
Library right now. I think I'm getting close to having that part working, though.
I feel like I've learned an awful lot lately. I've taught myself some reverse
engineering, learned more intricate details of how USB works, read the TLS spec and a good portion of the
TCPA spec
(pdf), and come to understand much more deeply how cryptography is
implemented. I've written a big integer library, starting with increment and
going to addition to multiplication to modular exponentiation, and deduced
several techniques that will make such code faster.
And that's just what I've done in my non-school time. I've learned how speech
recognition works, about a thousand problems that are equivalent to
least-congestion routing, and a similar number of ways to solve a linear system
of equations. I've learned that most security techniques aren't about
providing absolute security, but rather making things less easy for people to
screw up. I've found that satisficing is way more effective a strategy for
getting things done than seeking perfection, but that the latter's purity is
itself rewarding. I've started toying with an idea that I think I might be
able to turn into a thesis.
Yet, for every fact I learn, I find a dozen more questions, a pile of things I
still have yet to learn. There are a zillion lectures that I want to go hear.
I have a significant reading list backlogged. My TODO list above is but a tiny
sample of what I actually have to do. There is never enough time. I must
choose what is worthwhile.
I wonder how close I am to ten-thousand hours
spent working with computers. Maybe I've gone over it already. I hope I'm choosing wisely.
Recently, someone asked me through a mutual friend how to become pro enough
with computers to pass a particular (rather stringent) interview. I replied
with a link to Peter Norvig's
Teach Yourself Programming in Ten Years.
I didn't mean to be snarky, simply
realistic. My friend replied: "it is upsetting to think that I can only be
great at a handful of things before I die." It's true. That limiting factor
is enough to make you question the sincerity of someone who claims to want to
learn something. How much can they really want to understand? Should any of
us bother to be a mentor at our field of expertise for someone who isn't
willing to put forth the same effort we did to gain our talent?
Is it fair to expect such reciprocity? Is it just to demand it? Is it kind?
Who am I to judge whether you're trying to get the cheap benefit of my labor,
or if you too are beginning the journey of a lifetime? How could I stand to
crush another's curiosity, when curiosity and the passion to learn are the most
beautiful things I know in this world?
I will share everything I can, for that is what I would wish from everyone
else. In turn, I must recognize that learning doesn't always happen on my own
terms. I hope you'll do the same.
I'm behind on blogging. I have an extensive writeup of Monday and part of
Tuesday to do. They will be post-dated, but eventually posted.
Today, the BiD lab speaker was Nic Ducheneaut,
speaking about massively multiplayer online games; namely, the
interactions and social structure formed by players of World of Warcraft. This
talk attracted a wide array of new faces, including a girl who had hair down to
her legs. I wanted to tell her that her hair was awesome and impressive, but
had to leave for class. But I digress.
It turns out that WoW is a great platform for research, since there are
long-lived groups, piles of quantitative data, and consistency of the game over
different cultures (US, Europe, and Asia servers all have the same game
structure). It also turns out that the US players are more careless, have
less organized groups, and in general, seem to take the game and social structure
less seriously. The US players die more than those in Hong Kong or Taiwan, despite
spending less time in combat. Takeaway: the US is reckless.
Players tend to start out playing solo a greater percentage of the time, but as
they progress, they spend more and more of their time in groups. Part of this
is the way the game is structured - as tasks grow harder, you eventually reach
a level cap, so the only way to continue to increase strength is by adding
numbers to your group.
I found it fascinating that the majority of guilds did not survive more than 6
months, which suggests that organizing and maintaining a guild is a challenging
task (which isn't so surprising, really). Further, about 90% of players are in
a guild with 35 people or less. Back in the days of raids that required 40
players, this meant that game content was inaccessible to the majority of
players. Now, Blizzard has changed to 20-man raids, which helps make them more
approachable. The larger guilds have very military-like structures. I suppose
that's what they are, in-game.
Players also are moving toward indirect socialization, much in the way that
Facebook and Twitter are more passive forms of interaction than those of old.
Nic made no judgment of this interaction, simply a statement of fact. He also
mentioned Putnam's Bowling Alone,
which made me think back to that excellent political science class I took.
All in all, very fascinating work.
Now I'm doing my midterm, for real. Explanation will come in a post-dated entry.
This post will discuss events leading up to, stories from, and aftermath of The
Game, a scavenger puzzle hunt associated with Stanford. Yes, I'll get to last
Monday's post eventually.
A good while back, my labmate Lora
introduced me to (among others) her friend Soja, who was in the process of
writing a puzzle hunt, and was toying with Rubik's cubes. This Saturday (the
9th) was the date of The Game itself. Lora, Wes, and I formed a team
named PhDestroy! for The Game.
We were in the first round of competitors, so we were supposed to report to the
Hacker Dojo in Mountain View 20 minutes
before our designated start time of 09:30. Since that is early o'clock, and
since Lora's parents live in Los Altos (much closer to Mountain View), we opted
to meet up Friday night, make dinner, drive down to Los Altos, spend the night
there, and sleep in a little more the day of the game. Lora brought fresh
produce from her CSA drop which produced a salad and roast asparagus, and I
made the delicious pineapple sausages that I've swooned over before.
Om nom nom.
Saturday morning, we reported to the Hacker Dojo at about 09:15 (ehh, whatever)
to sign release paperwork. Costumes were encouraged, so I wore my trenchcoat
and my wizard hat. One of Lora and Wes sported a grey fedora - I don't
remember who wore it more, nor who started wearing it, but it worked for both
of them.
We had a veritable Smörgåsbord of other things we brought, in case they were
useful:
Pencil and paper
Cell phones (with Google Goggles)
A guitar (which Wes played with great skill, making our traveling hours splendid)
Laptops. Including the Cr-48 Google provided me, with its Verizon 3G. This was nothing short of FANTASTIC - it was incredibly convenient to have a full-size keyboard for quickly doing searches, writing programs, and exploring data.
An inverter for use in the car. This was also excellent, because our phones/laptops would not have lasted through the competition otherwise.
Headlamps (for when night falls)
Good puzzle-hunting attitudes
Rather than give a blow-by-blow of every puzzle, I'll just mention two of my favorites:
At the Gates of Hell on Stanford campus, we found a Rubik's cube labelled with
some numbers, red or pink hearts, and stars with what appeared to be Braille
dots on them. The numbers indicated which of the six faces a tile belonged on,
but not all tiles were labelled. I spent a good while working out the
constraints, labelling each tile and eventually solving the cube. Each face
had three star tiles on the right, and a pattern of red and pink hearts on the
left which formed a 3x2 grid, just like a Braille letter. After some fussing
about and pinging Game Control which told us there was a typo (the Rs should
have been Os), we figured out that the clue read AT EL PALO ALTO (not
ATELPALRALTR). I love Rubik's cubes, and (with the exception of the typo),
this puzzle was very well put together - it gave enough information such that
the most logical way to proceed was the correct one, but left enough to the
solver to still be fun.
A webpage with some ASCII art made of a bunch of letters from the set {A, G, C,
T}. I broke out the Cr-48, ssh'd into my server, and quickly wrote a Python
script for converting three-letter bases into their corresponding single-letter
amino acid codes. This was the point where I realized JUST how awesome having
the Cr-48 was going to be. Thanks, Google!
There were also several puzzles that we didn't get right off the bat, to say
the least. We spent about an hour trying to find a clue in the
plot-development page that contained none, rather than looking at the clue
given at the next site. We spent a rather long time before we learned to read
interleaved 2-of-5 barcodes
manually. This gave us a pair of GPS coordinates, which we assumed were in
degrees. By the time Game Control confirmed the clue with "yes, m/s" we were
already halfway to Saratoga. Whoops.
We also wound up severely off-course when we were trying to solve a puzzle
titled "Real irrational" which should have taken us two minutes if we'd
realized that we were working with octal numbers for ASCII characters, but
instead we wound up with four three-digit numbers that we tried to make sense
of by going to addresses on the El Camino Real (same name, right?). This wound
up putting us in Sunnyvale, which was also way too far southeast. On the
upside, we stopped and got food, which helped put us all in better moods.
As a team, we were probably a bit too stubborn with the puzzles. We didn't
know how long we should wait before asking for hints, or what the relative
penalty for asking for a hint would be, so being the self-reliant folk we are,
we generally didn't ask until we'd been stumped well over an hour. This,
combined with Game control being understaffed (and accordingly swamped with
calls), put us a bit behind the curve. I'd say our teamwork wasn't bad - Wes
provided interesting thoughts and interpretations (and awesome soothing guitar
music on the road), Lora had local knowledge and navigation prowess, and I
drove and put my mad computer skills to use where applicable.
We finished The Game at 02:06 Sunday morning. The endgame gives you an
administrative password that you're supposed to give to one of three
characters: a sentient computer program that plays the stock market, the
programmer who wrote it and feels it's gotten out of hand, and the
mafia-associated CEO of the company that formerly employed the programmer who
wrote the aforementioned program. Any choice was considered a win condition.
We made the unique decision to give the password to all three characters,
telling them they should work out their problems in a peaceful, productive
manner. This amused Game Control.
We returned to Lora's parents' house in Los Altos for the night, and I fell
asleep in less than ten minutes. I woke up due to sunlight around 8, did some
quick subtraction (8 - 2:30 = not enough) and went back to sleep until 11:45,
when I was awoken by a knock on my door and Lora informing me that "French toast
is best when warm." Awesome.
We showed up (about 45 minutes late; whatever) to the post-Game reception at
the Hacker Dojo, titled "Mimrsas and frrd" (a good-natured poke at the
misspelled Rubik's cube clue), and got a pretty decent rundown from Game
Control of how stuff happened. It was very comforting to know that some of the
issues that we'd had weren't entirely our fault. We all offered our (legion)
constructive criticism, and I think I volunteered to playtest the next game
they write. Maybe I'll help Game Control next time. We'll see.
All in all, I had fun, despite a few lows over the day. Thanks to Game Control
for putting the event on, and thanks to Lora and Wes for teaming up with me for a
great day of puzzle-hunting!
TL;DR I'm leaving Facebook on my birthday and deleting my account; please
follow my blog and email
(drew.m.fisher@gmail.com) or IM me (drew.m.fisher@gmail.com for Jabber,
drew.m.fisher on Skype) if you want to keep in touch.
I'm leaving Facebook. On my birthday, April 24th, I will delete my account.
Why? Several reasons:
First, it was a time sink. I read lots, and saw lots of generally useless
noise. I learned little from my time spent on Facebook. It's telling that I
came to think of it as time spent, rather than time well used. It was no
longer a good use of my time. It became filled with uninteresting fluff.
Furthermore, as an obsessive engineer, I seek the best tools for each task I
wish to perform. Facebook, while decent at a variety of tasks, is not the
best at any of them:
Planning an event? Try Doodle for cooperative planning, or Evite for a more unilateral approach. Or just email people.
Group conversation? Email, particularly with a mailinglist, is far superior to attempting to do the equivalent with Facebook messages. Far better message threading, and a pile of clients ranging from web interfaces like Gmail to rich clients like Thunderbird or Outlook. These let you organize your communications in a much more useful manner with a snappier interface to boot, which means time saved both now and later.
Chat? Pick any of a multitude of chat clients that you can use features like tabs with, or if you want it in your browser, imo (full disclosure: I will be interning with imo this summer). There's even iPhone and Android apps.
Sharing photos? Try Flickr or PicasaWeb, or anywhere that doesn't so horribly restrict the resolution of the photos you can share. Autotagging, resizing, cropping, and all the important image-editing features are both faster and easier to use in native apps anyway.
News? If you can't find interesting things to read on the Internet, something is wrong with you. Subscribe to some RSS feeds. Read your friends' blogs. If you have WAAAAY too much time on your hands, go to reddit.
Keeping up with what your friends are up to? Write at a higher grade level about your life on a blog, or if you just want the 140-character version, use your favorite microblogging site like twitter.com or identi.ca.
Interactive group chat? IRC chatrooms, if your IM network doesn't support them (most do).
All of the above offer better workflow for their own particular task, more
powerful features, and in the long run, more time saved. Few, if any, require
any more learning than the comparable tools in Facebook's own.
I came to the conclusion that pretty much the only thing Facebook had going for
it was that many people already had an account and spent copious time on the
site. While this is perhaps relevant for things that require site
registration, I found a very uncompelling argument to keep with it.
Finally, I take issue with Facebook's treatment of its users and their privacy.
I recognize that most of you probably won't find this very important or
compelling, but Facebook has a pretty long history of:
You all know I'm a techie. For a security- and privacy-minded individual like
myself, this sort of behavior (and ongoing trend toward the worse) is deeply
troubling. I've complained about these for quite some time. I realized that
if I wanted to have any moral authority whatsoever, I needed to put my money
where my mouth was, and stop using Facebook myself.
"But," I thought, "how could I leave Facebook? Everyone has an account! It's
the primary method a lot of people use to keep in touch with their friends!
Think of what you'd lose if you left!"
What would I lose if I left? As a scientist, I decided that the best way to
answer my question would be to perform an experiment. Last summer, starting
some time around July (I forget the exact date), I stopped logging in to
Facebook altogether. Well, with single-digit exceptions over 9 months. Close
enough.
It wasn't as bad as I thought it would be. I found myself with more time to
read fascinating things, work on personal projects, and enjoy my time in the
real world. I finished an electronics project that I'd been thinking about
doing for four years. I started writing a blog of the happenings of my life,
including writing the software that powers it. Many friends still instant
messaged me. I got pictures from friends by email. My mom continued to call
me via Skype and ordinary telephone. I got the most epic box of brownies I
have seen in my life in the mail. If anything, making the interaction more
scarce had made it more precious, and the time I gained was phenomenal.
I consider my experiment a success. I determined that I can live and live well
without Facebook. When Sarah Luna wrote that she would leave Facebook on her
birthday, I decided that I, too, would delete my account on my birthday. I see
it as simply making the gains from the experiment permanent. Moving to Phase
2.
In recognition of the fact that for some of you, this may be the only form of
contact by which you know to reach me, I am writing this note to offer you
alternatives. It is not my intent to sever contact with all you wonderful
people. I simply ask that if you wish to reach me, you do it through any of a
multitude of methods that do not involve Facebook:
Email: drew.m.fisher@gmail.com
Phone: 214-316-8685
GChat: drew.m.fisher@gmail.com
Skype: drew.m.fisher
Blog: http://zarvox.org/shortlog/
Contact me, and I will reply happily and at length (although I can't
necessarily guarantee promptness). If you write a blog, send me the URL, and
I'll subscribe to your feed. Comment on mine and I will write back. I do
still want to hear about your lives.
And perhaps consider: what would you really lose if you left, too? Could you
make better use of your time?
receiving numerous thoughtful comments on my previous post, in the form of Facebook comments, Facebook messages, emails, and IMs
discovering that someone, somewhere, managed to land on my blog from Google with the search query "You have 5 types of shocks, and you are late for work. You don't have time to look so you just randomly grab socks out of your drawer. What is the minimum amount of socks needed to ensure you got a matching pair". How, I do not know.
Ivan Tashev - designed audio pipeline architecture and DSP algorithms for the Kinect and has lots of papers on related topics
Ramani Duraiswami - research interests include lots of work on microphone arrays and acoustics work
Adam O'Donovan - work on beamforming and spherical microphone arrays; also apparently interned at Microsoft during the right time to have worked on Kinect
I find it kinda cool, learning the names of the experts in my various fields of
study. There's only so many, after all, and you start to recognize the names
and faces along with the work. That's something I hadn't expected - I guess I
never really though about the fact that the number of researchers in a field
was finite. My daily dose of "small world, huh?"
Among piles of other things, I'm in the process of seeking housing in Palo Alto
for the summer and Berkeley again in the fall. Today, I learned in great
detail about the concept of "soft story"
buildings, after seeing a sign at one of the complexes we visited that referred
to Berkley Municipal Code section 19.39.060.
It turns out that Berkeley has a program
for identifying soft-story buildings that might be dangerous in the event of an
earthquake. Since it's rather likely that we'll have a significant earthquake
in the next couple decades, and since lots of housing in Berkeley would not
survive such an earthquake, the city created measures to help protect the city
against the expected natural disaster: all buildings that were likely to have
soft-story weakness were required to have an inspection by a professional
engineer, who would file a report on the building's safety. If the building
was considered potentially hazardous, the building owner would be required to
make tenants and visitors aware through letters and signage that it was known
to be possibly unsafe, until it is retrofit with structural reinforcements to
bring the building up to a higher standard of safety. No mandate to actually
retrofit the buildings was given, but the action taken effectively incentivizes
landlords to fix their dangerous buildings by making them less desirable on the
housing market.
One of the apartments we went to see was on Berkeley's Inventory of Potentially Hazardous Buildings (pdf). I've pretty much reached the
conclusion that as nice as the apartment was, it's a pretty bad idea for me to
choose to live in housing on the Potentially Hazardous Buildings. My life and
possessions in the event of a
rather likely seismic event
are worth more than saving a hundred bucks a month.
I'd rather not have my car go like this.
In other news: I filed for my 2010 tax returns today. I seriously need to
start upping my withholdings exemptions, because this is a second year that
I've given the US government a sizable several-month interest-free loan. Then
again, goodness knows they need it.
Things that I have done to mitigate stress the past few days:
Watched the first episode of A Game of Thrones with the awesome folks at the house on Keith. I'm reading the book at the minimal rate needed to stay ahead of the TV series, because I don't have time to devour the book all at once right now, but I also consider it a crime to watch a movie or TV series based on a book without first reading said book.
Rocked out in the lab on keyboards with Andy on drums and Wes on guitar. That's right, BiD band is practicing our two-song setlist. Woohoo.
Joined Lora and Wes for a tasty dinner of homemade udon (thanks for cooking, Lora!). Afterward, Lora brought out her violin and put up with me stumbling through the piano half of a Sonata for violin and piano. Last week's session excepted, I hadn't played piano since winter break. My sightreading could stand improvement.
Napped. Sleep is a wonderful thing.
Called my parents.
Read through my server logs. Seeing that people are reading my blog gives me warm fuzzies.
The Word "Hacker" by Paul Graham. For the record, I like nearly every one of Graham's essays.
Amazon's been having some
rathersignificantdowntime
lately. I can't wait to read the post-mortem analysis, when it's all over. It
does make one wonder how worthwhile throwing everything to the cloud really is.
Even Microsoft
doesn't have this totally figured out yet.
Today, I joined a bunch of Stanford and Berkeley folk for competing in Google
Games, which is kinda like a puzzle hunt and trivia game and a coding
competition. I really do love these puzzle hunt activities. I hooked up a
bruteforcer to a slitherlink puzzle, coded one of the programming problems, and
had valuable insights on several puzzles. One of them, a word association
puzzle, was widely hated by all the teams I talked to, but I rather enjoyed the
rest of the competition, even if we got quite stuck on silly things a couple
times.
Coincidentally, I ran into several folks I'd known before:
Kat Busch, who I'd met exactly once - two summers ago, Sarah and I had gone to Stanford to visit her friend Evelyn, and we managed to go catch the end of Evelyn's boyfriend's belt test, and then afterward we all went for sushi in downtown Palo Alto. Fun!
Edward Schmerling (a hilarious character as usual) and Patrick Thill, whom I'd teamed up with several years ago for the Texas ARML team. Sharp folks. Their team took second.
It's a small world.
After the competition, our team visited Palo Alto/Stanford (decked out in all
our Berkeley gear...this produced some looks and some interesting
conversations). We visited the Gates building, and saw the Gates of Hell, and
passed by Steve Jobs' house, and the HP garage. Stanford is gorgeous, and very
spacious.
Then, we broke out our smartphones and found a pizza place in SF where we had
some deep-dish pizza. It was tasty, but a little heavy on the tomato sauce. I
prefer Sicilia's, but that's way on the other coast. Alas.
This year, I moved to a new place, and made many new acquaintences that are
turning into friends. Friends that I anticipate growing closer to and enjoying
for many years to come. Friends that will see me at my best, and at my worst,
and will see to it that I pass through both. Friends that will challenge me,
guide me, and teach me. Friends whom I hope I am teaching in return.
I started a blog. I closed my Facebook account. I joined a community through
which I (re)awoke a passion and made friends in at least four countries.
Hopefully I'll have the pleasure of meeting them in person some day. I started
making significant contributions to the open-source community. I lived for
nine months in an apartment, and made my own food, and didn't starve to death.
I woke up this morning around 08:00 to the deafening sound of silence. Most
ordinary people would enjoy the peace and quiet, but even in my pre-alarm
stupor, I knew that something was wrong. See, I leave my computer on 24/7, and
if it's silent in my room, that means my computer isn't running. And when my
computer isn't running, I have a problem.
I rolled over and saw the power light on the front of the tower blinking amber.
An SMS on my phone from my system monitoring solution confirmed "kraken is
DOWN." (Yes, I name my systems after mythological creatures.) A look at the
repair manual
(good thing I've got phoenix, my laptop!) told me this means "The computer is
receiving electrical power, but an internal power problem might exist."
I opened up the case, and after a brief bit of looking around, found this:
That's a blown capacitor on the motherboard (the one with the golden stipe);
while the lighting is nothing great, you can see the top is not uniformly the
same color as the rest of the capacitors.
kraken has served me well (and nearly constantly) for 7 years now, so it's not
like this is terribly surprising. Nor is it a particularly damaging failure -
my disks are perfectly fine, and my data is backed up to phoenix anyhow. It
is, however, a nuisance to deal with this time of year.
If I had a comparable capacitor on hand, I might try soldering a new one in its
place, but 2200μF at 6.3V capacitors are a part neither I nor the BiD lab have
lying around.
So, after 7 years of excellent use, it is time for me to get a new computer.
I'm still deciding what to get, and would appreciate suggestions of:
Cases. I'm looking for something quiet, not too heavy, and easy to work inside. Due to the way I use my desktop, I do not want/need tons of shiny bright lights - these would make it harder to sleep at night.
Power supplies. SATA power is a must
Motherboards. I need at least 4 SATA ports and at least one IDE channel. Dual gigabit NICs would be excellent, but I have a PCI-e 1x gigabit ethernet card already if not. Plenty of PCI slots would also be excellent, since I have an ath5k wireless card and a sound card as well, though I don't mind dropping the soundcard if the onboard audio is better anyway (which, frankly, it should be).
CPUs. Something that could potentially last me about as long as this P4 did. Hardware virtualization extensions required. I'm open to both AMD or Intel offerings, but appreciate justification of suggestions from either side.
I opted to wait on purchasing a replacement desktop until my new credit card
arrives, so I can use the purchase of the new computer to hit the card spending
requirement (landing me some number of airline miles), then pay it off and not
use it again. Either way, I'm aiming for ~$1k, which means I'm probably going
to buy a fairly high-end Sandy Bridge processor, a ton of RAM, a motherboard
with dual NICs, and as silent a case as I can find to put it all in.
I talked to David Sun in lab today about his work involving localization with
array microphones, and he gave me some pointers that might help in implementing
an open-source sound localization library for use with the Kinect. Most of
that work is currently on hold, though, because 1) all my notes on the
cryptographic handshake are on my desktop, safely saved, but currently
inaccessible without yanking the disks and doing some magic with external
housings that might upset the RAID1, and 2) I HAVE PROJECTS THAT ARE DUE SOON
AAAHHHHH!!
On that topic: our security term project is progressing. Preliminary tests
were successful and show a 10x slowdown to network traffic with our technique,
which may turn out to be acceptable. Promising results, at least, which we'll
detail in our poster session on Friday. TODO: set up printing on the giant
plotter in the lab.
Today's poster session for computer security class went well, despite running a
bit long. It was scheduled from 4-6pm, and as of 18:15 our group still hadn't
started presenting to the professor. There was also a small mishap in printing
the poster yesterday, where we couldn't remember poster sizes and printed a
4'x3' poster (which turned out to be GIGANTIC) before realizing we probably
were supposed to make a 3'x2' one. OOPS. We taped the giant one to the wall
and went with it anyway.
Tomorrow is the Berkeley Mystery Hunt, and I'm SUPER excited for puzzles.
Here's to hoping it's well-written, well-run, and well-played. :)
Yesteday was the first Berkeley Mystery Hunt! I captained a team by the name
of PhDestroy! which included Lora and Wes (with whom I teamed up
for Stanford's The Game) and added Andy from the BiD lab,
Nima (with whom I joined forces for the College Puzzle Challenge),
one guy Nima knew named Di, and one guy named Bobby from the mailinglist.
Our merry band of 7 puzzlers got off to an incredibly slow start, solving
something like one puzzle in the first four hours, but as we got more hints and
worked out more, we started picking up the pace. I won't talk about the
puzzles, since the organizers want to run this hunt again for a Bay Area Night
Game at some point, but suffice it to say that we eventually got our act
together, solved the metameta, and at 10:53pm, won the game. (Actually, HQ had
to cut the entire runaround because we were the first team to get there, and
they wanted to finish by 10:30, with wrapup at 11. Oh well.)
Amusing events involved Andy trying to wheedle puzzle solutions out of Hunt HQ,
getting to reuse a piece of Python I've had on my computer for over three
years, and exchanging stories with the organizers.
I've volunteered to help the organizers write next year's hunt, and have some
good ideas for puzzles floating around in my head. That's another major step
toward completion of an item on my bucket list. I do hope I'll have the time
to help make this happen - I really do love these puzzle hunts. :)
One special thanks to my team, who made that day delightful (even when we were
stuck and frustrated) and one to the organizers, who put the whole thing together
over the course of several years. I had a blast!
UPDATE: photo! From left to right: Bobby Liue, Di Wang, Drew Fisher, Nima
Ahmadi Pour Anari, Andy Carle, Lora Oehlberg, Wes Willett. Thanks to Ankur
Mehta for the photo.
I am hopelessly behind on my blogging. In my defense, it's the end of the semester and I've been busy.
Monday: BiD took a field trip down to Menlo Park to visit Willow Garage. As
expected, it's a darn cool place. We got to tour the facilities, see all the
robots and demonstrations, and ate a delicious lunch. I chatted with
Radu Bogdan Rusu about object tracking.
Afterward, I showed off the Kinect audio driver I've written to Nate Koenig,
and he seemed to think it was cool.
That evening, the toilet in the half bath broke. Specifically, the shutoff
valve had broken, so the bowl refill tube was constantly spewing water.
Ordinarily, this would be a small problem, except for the fact that the metal
stop valve at the base of the toilet was so stuck that neither Andrew nor I
could budge it, even with the assistance of a wrench and a power screwdriver.
Thus, we had a pipe spewing water and no way to shut it off. Andrew jury-rigged
the toilet to be in a constant state of flushing so it would drain. Yay for
engineering!
Tuesday: I went to the bank to retrieve my passport from the safe deposit box.
The plumber came by, but couldn't shut off the water until the afternoon.
I also introduced Sarah to the concept of airbnb,
a site for finding rooms/couches/airbeds to crash on in various cities. It's
the appropriate tool for the slightly scrappy minimalist who'd rather find a
more inexpensive room with friendly people than stay in a hotel.
Wednesday: my panic level escalated as I prepared for my final presentation for
my algorithms class. I also picked up my midterm, which was absolutely utterly
a mistake.
Thursday: finished preparing my presentation and delivered it (after much
searching about for an empty space in which to give the talk). I had the
misfortune of having to present after two theory students who had WAY more
relevant-to-the-class talks than I did. But hey, that's how it goes. It'll go
the other way when I get to TA the HCI class next fall and everyone has to code
stuff. :P
After dinner, I called my dad and talked for a while, then went over to
Kristin's house to assemble her new desktop. We were able to get everything up
and running with few troubles - one too few power cables, the unfortunate
combination of an IDE DVD drive and a motherboard with only SATA ports, and an
upgrade-only Windows installer gave us a little grief, but all hardware worked
correctly, and with a USB flash drive we were successful in flashing the BIOS,
(double-)installing Windows, and getting everything up and running. Now I can
say I've assembled a computer from parts!
Today: I packed my luggage this morning and am flying to Vancouver for
CHI (which is why I had to present my project on
Thursday). I'm on the same plane as Wes and Kenghao from BiD lab, so it should
be pretty fun. I'll see if I can find a prepaid SIM card or something while
I'm in Canada; I don't really want to pay roaming charges.
UPDATE: I've arrived in Vancouver, and the view from the apartment we're staying
in is AWESOME. It'll be even better once the fog clears. Pictures will come later.
Wow. I have had an AWESOME week at CHI, and I have LOTS to write about, but I
also have a final project paper to finish writing up today, another paper that
we're supposed to be submitting on Monday, and I'm moving to Palo Alto on
Sunday to boot. So (sadly) stories will have to wait.
Oh, and I'm back in the US safely. In case you were wondering.
Update (rewrite?) from two weeks after the fact: Got up rather early and caught
the train back to Vancouver Airport. At the terminal, I ran into Anand
Kulkarni and Truc Nguyen, who happened to be on the same flight back to SFO.
After getting home, I ate some food (I honestly don't remember what), crashed
for about 30 minutes, and then went to campus to join Edward Wu and Justin
Samuel in writing up our research paper for our security class. It got a
little rough at the end, but we submitted our paper precisely on time and I
delivered a hardcopy of the finished product to the box ouside Dawn Song's
door. Then I went home and slept.
I spent basically the entire day in the BiD lab (or elsewhere on campus, like
in the 5th floor conference room in Soda hall, because the lab gets COLD)
working with Keng-hao on our paper submission for the Wireless Health
conference. By the end of this day, I was quite sick of writing papers.
Today I moved to my summer residence in Palo Alto. Well, for some value of
"moved." I didn't bring much: just my bedding, towel set, clothing, laptops,
and Kinect. Getting the queen-sized mattress into my Avalon alone was rather
tricky - I ended up using the orange luggage strap that Bloomberg gave me at
CHI to keep the mattress folded in half in the long dimension, and somehow
managed to stuff that unweildy mass in the back seat. The floor of the parking
garage was rather dirty, and it had been raining, so the mattress cover got a
bit dirty and damp in places. I'm washing it.
The drive was uneventful - I left ten minutes later than I had planned, but
Google Maps had overestimated travel time by the same ten minutes, so I arrived
right on time. Robby and I did a walkthrough of the house with the woman who
gave us the keys. It's quite a nice place - two-story, six bedroom, three
bathrooms, sizable kitchen, lots of downstairs common areas, washer and dryer,
backyard...all in pretty good condition. It's about 4 miles from the office
I'll be working at this summer.
After moving my stuff in, Robby and I moved a good deal of his possessions from
his previous residence to the house, with an intermediate stop at the IMO
office, where we set up my computers. I have a desktop running Kubuntu and a
laptop running OSX. This should be fun.
I ate dinner at the Korean BBQ restaurant in downtown Palo Alto for dinner; the
same place that Evelyn and Sarah and I visited two summers ago. It was
delicious.
Now I'm wishing I'd had room for my chair in the first trip from Berkeley. I'm
sitting on the kitchen floor with a $9 IKEA table in front of me, bundled up in
the blanket my sister gave me. I'll be going to bed pretty soon, methinks.
We still won't have internet at the house for a while. Thank goodness I can
tether with my phone. And by that, I mean: thanks, Nokia.
Since I'm a computer scientist, yesterday was my zeroth day of work, which
makes today my first day at work. It took a while to get everything up and
running, but the last thing I accomplished before going home was getting the
product running on my desktop. I considered this a reasonable accomplishment
for the first day, given the complexity involved. Hopefully my experience will
help get the other interns up and running even faster.
I made my first commit to the IMO source tree today.
Startups are awesome - you move fast. This one is particularly awesome -
interesting problems, a product that I rather like, and a fantastic work
environment. Dual 30" monitors, food/drinks all the time,
KDE on my desktop...it's pretty sweet.
Jeez, I am behind on my blogging. I still haven't written about CHI, nor my
papers, nor my adventures in Palo Alto thus far (nor my reverse-engineering
adventures, in case you had forgotten about that writeup. Still on hold).
It's all been pretty excellent, but pretty busy. Sadly, I need to drive for
about two hours tonight still, and I lack speedy Internet at my place in Palo
Alto (I'm still tethering with my phone, and getting Comcast to listen to us is
another story still). If it's not too late when I get back, maybe I'll paint
the tales I've been mentioning thus far in greater detail.
Update from 2011-05-27: a real post:
There's a decent contingent of IMO employees who like rock climbing, and the
indoor climbing gym they usually go to has this deal where they can refer a
friend, and the friend gets a free lesson and day of climbing, so I tagged
along. Somehow we crammed five people into Tom's tiny sports car. They showed
me the ropes, and I already knew the knots (hooray, Boy Scouts), so shortly
thereafter, I joined the others to climb some routes. It's much more planning
and finesse than I had thought. It'll be a while before proper form becomes
innate, but now I know: stay close to the wall, keep your arms straight, and
your legs bent. I climbed a 5.3 during the lesson, a few 5.7s, and a couple of
5.8s. My arms eventually reached the point at which I can no longer grasp
anything, let alone hold my weight (which, of course, I'd need to do less if I
had proper form). My veins were sticking out of my arms in a very foreign way.
It was awesome.
Nicolay took us though a long sequence of stretches afterward. I'm fairly
certain I have that stretching session to thank for my arms not falling off the
next day.
After the gym closed, John Rizzo and I agreed that we wanted food, so we went
to In-N-Out and had tasty burgers. They were delicious. I got home at some
terribly late hour in the morning, but it was an awesome day.
This evening, a group of 6 of us IMO folk went to
dinner at the Cheesecake Factory. While waiting to be seated, Robby and John
discussed how I should order the "2700." I hadn't seen such a dish on the menu
before, but it turns out that 2700 refers to the number of calories in the
dish.
After we were seated, and we got to look at the menus, we found that the 2700
had in fact been upgraded to the 2990. I opted for one of the 1770's instead,
which was delicious.
I sat on the couch and spent many consecutive hours reading A Game of Thrones.
Andrew and I set up our mini-grill and stuffed ourselves silly with grilled
sausages, grilled pineapple, and grilled bell peppers. DELICIOUS. I drove
back to Palo Alto and met the third bro of Brohaus (who moved in on Saturday).
Oh, that's our business' name, as far as Comcast is concerned. Having Internet
access is a wonderful thing.
So, the parts I ordered for my desktop have mostly shipped, with the exception
of the case that I would be putting everything in, because the place I ordered
it from ran out of stock. Stuff will arrive Thursday, but I probably won't be
able to build my machine until the new case comes in (which could be another
while). I still have to decide what case I want. I was aiming for a quiet
case without too many glowy lights - they're pretty, but annoying at night
and when you're trying to watch movies and stuff. On the other hand, if I
had more light from my desktop case, I might not have to buy a lamp for my
room in Berkeley. Decisions, decisions.
You know you like your job when you spend over 14 hours at the office and don't
even notice when it's past midnight and you're still fascinated by a problem.
I rather like my job - we have some crazy problems, but we have a bunch of
sharp folks, and I have all the tools I need. I am the limiting factor in what
I accomplish, and that is a very empowering feeling.
I plan to make back-dated posts for much of the past few weeks. The CHI posts
will probably take a little longer since they're less clear in my mind and I
have pictures, which means that I have to decide which ones to post and format
and resize and rename and link them all, so even with scripts, it takes a
little longer to write the posts. To the people who would propose objectively
and technically better blogging solutions: LA LA LA I CAN'T HEAR YOU OVER THE
AWESOMENESS OF MY BLOGGING ENGINE. <3.
My new power supply for the desktop I'm building arrived today. The rest of
the parts should arrive tomorrow. I'll still be short a case, though. I've
operated without my desktop for over a month now. It feels kinda weird - I've
adapted to using my laptop and phone instead, but it's definitely put a damper
on my computer usage at home.
On the OpenKinect side of things: lostincake
just sent in an updated C# wrapper, so I'm going to test that out and hopefully merge/push
the changes as soon as I can verify that they work right. :)
I slept in. I went for a run in my Vibrams in the afternoon, and it was
beautiful. For the first 8 minutes, that is. Then it (very slowly!) started
raining. I ran home. Shortly thereafter, Robby and I drove to Wal-mart to get
assorting house things - dish soap, a dish scrubber, and trash cans. We also
stopped by BevMo! so Robby could stock up on BAWLS.
A successful venture, followed by me fooling around with Skype and trying to
get the C# libfreenect wrapper working.
I suppose it started out last night, when I put my sheets in the washer, got
distracted, and failed to move them to the dryer. Bit of an unpleasant
surprise when I meant to turn in for the night and my sheets were still wet. I
put them in the dryer, curled up in my blanket on the floor, and wound up
crashing until 6am, at which point I woke up, removed my sheets from the dryer
and made my bed, got in my freshly-made bed, and failed to fall back asleep for
the next two hours. I did hear a veritable symphony of birdcalls though.
I finally got the C# libfreenect wrappers working this morning (I think I had
to install a bunch of 32-bit compatibility libraries), so I pushed those
changes out.
Around 10am, Robby and his former housemate and I went to U-Haul to rent a
truck to move Robby's remaining stuff from his former residence to the Brohaus.
I then started crawling the Craigslist "free stuff" listings for nearby items
of interest. It turns out that someone had made a posting about a week ago
regarding an upright piano. I emailed the lady, and it turned out that the
piano was still there. We needed breakfast still, and I had just gotten an SMS
from Jono, who was visiting assorted family
members in Palo Alto, so I invited him to join us. We picked him up and had breakfast
at Hobee's. I had huevos rancheros, and some spectacular blueberry coffee cake.
After breakfast, we dropped Jono at the Caltrain station and picked up the
U-Haul truck from where we had parked it in Palo Alto and set off to obtain the
piano. It was not terribly far, but it was terribly heavy. It's a good thing
there were three other guys at the place we were picking the piano up from in
addition to our three, because it took some serious effort to get the thing out
the door and down the stairs. It took Robby and James and I an additional dose
of sincere effort to get the piano up the ramp into the truck.
Then we realized that we had no way to tie the piano down in the moving truck,
but we'd sooner find out how bad a thing that was than have to unload the
piano, pick up some gear, and reload it. The next 15 or so minutes were some
of the slowest, most frightened driving I've been party to in quite some time.
At every stop sign, the piano would roll around in the back, slamming into the
walls loudly as we slowed or accelerated. And Palo Alto has lots and lots of
speed bumps and stop signs. Let's not forget that you don't want to have to
make unprotected left turns. We managed to plot a route that had only one left
turn, and that was at a stoplight. After we finally arrived at the Brohaus, we
opened the back to survey the damage. The baseboard had fallen off the piano,
but the rest was (somewhat surprisingly) intact. We wheeled that sucker into
the garage, because the three of us lacked sufficient manpower to lift it up
the steps at the front entryway. Needless to say, the piano needs tuning.
After the piano adventure, we waited around at the Brohaus for a while, hoping
that the parking situation at the other house would improve. I took a nap
(which I felt quite entitled to, after last night's sheets error). Chuan
stopped by, accompanied by his friends Jason, Brad, and Jackie. With seven
people, we were able to move the piano properly indoors.
We sat outside in the backyard for a while, chatting. Around 5:20pm or so, we
decided that Robby and James and I would go back and pack lots of stuff from
the old house and meet Chuan and friends at The Counter for delicious burgers
at 7. Packing was mostly uneventful, although Robby's desk had to be removed
from the building through the window, since it wouldn't fit through the doorway
and if we disassembled it, the wood would splinter and not be fit for reassembly.
We finished packing stuff right around 7, drove the truck back to Brohaus, and
picked up Michael (who had just gotten back from SF) for our trip to The
Counter, which was merry and delicious.
Grilled pineapple is excellent on burgers.
This has been the first year I've actually consciously noticed the seasons
changing. I attribute this in large part to my lack of indoor lighting. I
can't stay up as late as I used to, because the constant darkness makes me
sleepy, even in the light of my monitor. I have a giant window on my wall,
which makes it impossible to ignore the sun's chipper rays in the morning, even
though the window is on the west side of the building. The effect is even
stronger in the Brohaus, where I have an eastern window. I am no longer able
to sleep in late unless I am overwhelmingly tired.
Perhaps another part of it has been the heater varying between functional and
off, over the course of the winter. You notice the cold more when you have no
choice but to face it. I wonder what else I'll notice in this manner.
...are for people who don't know how to use USB and netboot installers. ;)
The build took a while, since I'm still pretty inexperienced at building these
things. The fact that neither the case nor the motherboard had assembly
instructions didn't help. It's a darn good thing the one I built with Kristin
a month or so ago did - I probably would have missed a step or two otherwise.
The only two tools I needed (and technically, I could have done it without the
awesome screwdriver, just more slowly):
The case is one of the Antec ones that puts the power supply in the bottom of
the case. I'm still not sure how I feel about this. On the upside, the power
supply creates far less turbulance around the processor at the bottom, and the
case can have a vent at the top, which is more efficient since heat rises. On
the downside, it puts the power supply in a different place than most
motherboards and power supplies expect you to put such things, so some cables
may be too short to route comfortably, which can cause turbulence and reduce
cooling efficiency. Also, it's harder to flip the switch at the bottom than at
the top.
The case also had these cool pull-out bays for the hard disks with these huge rubberized shock absorbers:
All assembled (yes, the cables are a mess):
About five minutes of fiddling with UEFI BIOS settings, and I successfully
booted kraken. Later, I'll have to reinstall everything with 64-bit
packages, and set up VMs, and...I'll be busy this weekend, I imagine. :)
And yes, the thing is as quiet as I had hoped. With the case fans set to low
speed, I worried for a moment that the system didn't start when I pressed the power
button. Even at high speed, the fans are peaceful. I am quite pleased.
Cool day. I went to a SIGGRAPH meeting
featuring a talk by Johnny Lee (HCI
researcher and one of the core contributors to the Kinect project at Microsoft)
that I had heard about from qDot
(the other OpenKinect maintainer) three hours before the event. Thanks to the
fact that I'm already in Silicon Valley, I was able to go anyway.
It was cool hearing about how Microsoft had to learn how to deal with taking
research where things are often poorly or underspecified and being able to
integrate such a team with their software engineering prowess where they have a
lot of skill at successfully solving well-defined problems. One thing Johnny
noted in particular was the fact that traditional software engineering expects
linear progress over time - you expect to accomplish twice as much work in two
months with the same manpower as in one month. That doesn't hold true for
research - it's much more a series of quantum leaps, and much harder to predict
future performance. Now that I've heard about some more of the difficulties
involved in the Kinect project, I'm even more impressed by what Microsoft was
able to deliver. Congratulations, Johnny - you accomplished your goal of
commoditizing depth cameras. :)
After the talk, I hung around and chatted with the other attendees and Johnny
about all manner of HCI-related things until about 10pm. When I got home,
there were 20 snails in our front walkway. I counted them.
First story: I spent a decent portion of yesterday and today installing Windows
7 on kraken to make it a dual-boot machine. This was a longer and more
involved process than is ordinarily the case, and one which was made much worse
by my big mouth and stubborn attitude.
I'll start off by noting that two posts ago,
I said something to the effect of "optical drives are for chumps who don't know
how to do USB and netboot installs." So, obviously, when it came time for me
to do a Windows 7 install, you can bet what piece of hardware I wasn't going
to use.
I should note that two of my housemates offered their SATA optical drives, but
I felt obligated not to use them after talking big about how I didn't need one.
Windows 7 can be installed from a variety of bootable media, and in particular,
installing from a USB drive is officially supported. You just copy the data
from the DVD onto the flash drive, run G:\boot\bootsect.exe /nt60 G: (if G: is the drive letter of the
USB drive you just copied things to), and then boot from the USB drive. Magic.
So I used my laptop to rip my properly-purchased Windows 7 Ultimate disc into
an iso, and then I scp'd that image to my desktop. It was at this point that I
realized that I had left all my flash drives in Berkeley, in addition to all my
external hard disk drives. Crap.
I came up with a solution in fairly short order - I have a 16GB microSD card in
my Nokia N900 which had no data of importance on it (I was using it to play
around with MeeGo). The phone will present itself as a pair of USB disks when
hooked up to a computer via USB. So I would use my Linux phone to orchestrate
this Windows installation.
Since I refuse to run Windows programs as root in wine (it's just a bad idea),
I'd need a Windows install to do that
I have a 32-bit Windows XP virtual machine that runs inside
KVM. It is a relic from a different age, and I
hadn't used it in a while. I spent three hours bringing it up to date.
I could have used the VM to partition and format the memory card, but it turns
out that KVM only supports USB 1.1 for USB passthrough, so that would have been
terribly slow. On the upside, ntfsprogs
provides mkfs.ntfs, and I'm quite familiar with fdisk by now, so I was able to
format my "flash drive" and copy all the files from the DVD image to it at high speed
from Linux. Huzzah!
Then, I booted up the VM, tried to run bootsect.exe, and discovered that bit
with bootsect.exe from above has another limitation. In particular: the
Windows install running bootsect.exe has to be the same bitness as the Windows
7 installer image. That is, if you're using an installer DVD for 64-bit
Windows, you'll need to run bootsect.exe from a 64-bit windows install.
I don't have a 64-bit Windows install. But I did have some foresight,
some resources, and a lucky guess.
At work, several of us had needed to install Windows 7 Home Premium. Each of
our computers had come with a license, but the first thing we did after
unboxing them was install Ubuntu on them, so we put the Win7 installs in VMs.
To speed along the installs for myself and others, I dropped an iso of both the
32-bit and 64-bit installer discs. So, I ssh'd into my work computer, mounted
the 32-bit image loopback, scp'd thatbootsect.exe through three
intermediaries and into my WinXP VM, and ran it.
It worked.
So I booted the Windows installer from my Linux phone, ran through the install,
then ran through the install again (since this was the upgrade disc, but a full
license). Many hours later, I had a Windows 7 Ultimate install.
Then I used my laptop to put GNU GRUB on my phone, so I could use it to boot
Linux, so I could reinstall GRUB on the master boot record, so I could actually
dual-boot. Victory!
Second story: Georges Harik invited all of the IMO
employees over to watch the Dallas/Miami basketball game on his 3D TV. He has
one of the Samsung 3DTVs and something like 27 pairs of active-shutter glasses.
The effect was pretty neat. It was most appreciable when the 3D presented
itself as simply adding depth to the image - no objects in the frame "closer"
than the screen itself. When things appeared to float it midair, it just
looked weird. The glasses made my eyes kinda tired, and they were a little
tight around my skull. Sarah says my head is getting to big again.
Third story: I'm reading Surely You're Joking, Mr. Feynman!. Entertaining.
I've spent 10 hours in the datacenter over the past two days, unboxing,
racking, and configuring a bunch of new servers. Lots of things that were
supposed to work didn't. I did get to break out my sysadmin skills, though.
Unfortunately, in unboxing all the pizza boxes, I managed to cut a fairly deep
slice into my left thumb with my beloved Leatherman Skeletool. I suffered a -1
penalty to DEX for the rest of the day.
I find it oddly fitting that today's xkcd is
also about setting up servers. I've been staring at debian-installer prompts
for quite some time today.
Robert and I played Tetris Attack
(which is nothing like Tetris at all) for a couple hours after work today. I
went from fairly awful to winning between one in three and one in four games.
He wants us to train for and enter a Tetris Attack tournament. We shall see.
Saturday: I drove up to Berkeley and enjoyed delicious foods with the rest of
the cooking club at the house on Keith Street. There was spinach salad with
walnuts and strawberries, sliced pineapple, and TONS of delicious lasagna. I
had three slices. It was excellent. There was a cocunut cake with an amusing
slant - one of the pans had tilted in the oven, so one layer was a normal
cylinder shape, whereas the other was...much taller on one side than the other.
Tasty, either way.
The conversations were entertaining as usual. Later, we split up into two
teams of 5 and played Taboo. Justine and Jon Long and I are pretty awesome at
Taboo, if I may say so - we won after collecting some 13-point lead after an
equal number of turns. Perhaps part of this was Greg trying to answer "MYYYYYY
DIIIIIIIIIIICK" to every description. One particularly funny exchange went
something like this:
Me: Infinite mass, dark. (well, that's technically wrong, but the point was made)
Jon: Black hole! (Correct.)
Greg: MYYYYY DIIIIIICK!
Me: Well, it does have infinite mass.
Justine: And has never seen light.
(laughter until tears)
It's always a fun time with that crowd.
Sunday: I (along with Lora and Wes and Mark) helped Andy Carle move from his
place in El Cerrito to his new place in Richmond. Moving large pieces of
furniture is, as usual, a constraint satisfaction problem. One of the other
guys at the apartment complex had left his underwear out drying on a plank of
wood laid across the fence on either side of the walkway between the stairwell
and the curb. This made for an unusual obstacle when moving mattresses and
boxsprings and other furniture.
After we finished the move, we ate delicious pizza on the deck of the new
house. There were a few fruit trees that I'd never heard of before with these
small succulent fruits with huge seeds. Nominally tasty, though bees nearby
scared me. I guess I'm still not over my stinging insect phobia.
I drove back to Palo Alto, and Chuan had moved in and Michael had returned from
his travels. Brohaus is up to 6 bros now. The energy is excellent - I had
forgotten how much I liked living with tons of people. I long to return to the
dorms at A&M and MIT.
Our house has four ethernet switches forming a minimum spanning tree.
Monday: Wildly uneventful.
Tuesday: Before work, I called my mom. She's spent the past several months
planning all the details of this giant trip to Boston for her church that they
leave for tomorrow morning. I hope it goes well.
After work, I chatted with David Wagner on Skype about possible research
collaborations. Fun stuff - I may have more projects to join come Fall.
Wednesday (today): mostly uneventful, but made some good progress on my project
at work. When I got home, I googled one of the other imo interns and found his
webpage. Turns out he's an excellent pianist.
On that topic: I need to get a tuning kit for this piano that I have at the
house, so I can get it in working order.
Gah, birds are loud. It sounds like an aviary outside my window in the
morning.
Last night six of the guys in our house (myself included)
draftedZendikar. It was my first
time actually playing MtG proper - I'd watched others a fair bit, and
understood a decent amount of the mechanics, but I knew nothing about this set.
I wound up with a Blue/Green control/landfall deck that wound up winning 2-0
against Robby, 2-1 against Michael, and 2-0 against Chuan before people had to
go to bed (we'll finish the round-robin tonight or something). After that,
Chuan and I played exhibition matches until one of us reached 4 wins, which was
quite exciting when Chuan won the first three but I came back.
Today I'm meeting with Leila Takayama and Brian Gerkey of Willow Garage to
discuss the future of open-source Kinect drivers.
I read this essay and was so taken by it
that I felt the need to reshare it. The line that really spoke to me was in the
title:
There's no speed limit.
It's how I've felt my whole life about everything. As a child, I could not
understand why other people did not learn things at the same pace as I did.
It's why I read about as many things as I can, to this day.
The problem is, it asks what becomes an uncomfortable question - namely, are
people actually all created equal? If we accept that they are, then I am still
convinced that anyone can set an inordinate pace for themselves, and be able to
keep up with it, should they so choose. Really, there's no clear upper bound
to how fast you can learn something, if you don't worry yourself with matching
speeds with everyone around you. But that's not quite what we see happening in
the real world. And besides, it doesn't take the other half of the equation
into account - the intrinsic motivation of the learner. I've seen people who
didn't care about anything excel, and people who struggle with all their might
and fail.
There are two possible explanations that I can come up with, and perhaps the
answer lies somewhere in between. The first is that some people are simply
smarter than others, have more potential than others, are imbued with more
drive than others, and deserve more investment than others. But that's not
politically correct. No one wants to hear that their child is ugly, or lazy,
or a bully, even if it's the truth. To me, it seems as though people do not
value the truth as much as they value feeling good about themselves.
The other explanation I can offer is that our society is structured in a way
that discourages our young people from taking their education into their own
hands.
We learn from an early age that we should stay in line with our peers. If one
student pushes herself to finish an assignment early, she is told to sit and be
quiet until the other students are done. She implicitly learns that there's no
point in trying to learn as quickly as possible; she should simply stick with
the rest of her class. If she doesn't, it makes it hard on the teacher, who
unfortunately has to serve as both tutor and authority figure. And that
doesn't even touch on the social issues that arise when she's demonstrably more
productive than her peers. She's made into the common enemy of the other
students. What a terrible reward for being smart and proactive about her
education!
The problem with education with no speed limits is that such a system doesn't
scale the way we need it to to work with the society we've built. You can't
put 30 kids the same age in a classroom together to be supervised by a single
teacher if they aren't going to learn things together at the same pace. You'd
have utter chaos. No speed limits worked for Derek Sivers (from the link above)
because he had private lessons, where his teacher could adapt things to his
pace and set what others would call incredibly high expectations.
Education scales when you can have a few adults keep kids in a school all day
to keep them out of the way of other adults trying to get work done with the
hope of imparting some knowledge upon them. Don't get me wrong; I think that
education is important to having a developed and educated society like our own.
We'd pay a different productivity cost if we took the master-apprentice route -
we'd have a lot more people spending time teaching their trade to the
apprentices rather than putting it into practice.
The question is: is this worth it? Are there ways we can improve our options?
The most obvious way to improve the efficiency here is to push more
responsibility onto the excess resource in the reaction - the children. If the
learning is student-led, then the teachers don't have to be as numerous. The
kids will teach each other, too. Sugata Mitra's TED talk shows how it works.
I find it interesting that so many people that I know say that they didn't
learn how to learn until they reached college. Even given the limited sample,
that's a sobering thought - we're failing to teach the most important
foundations until we've passed through the vast majority of our formal
schooling.
Don't teach kids things. Teach them how to learn, and set the expectation that
they do so in abundance.
Great weekend. (And yes, I know I haven't blogged in a while. I've been busy.)
Saturday: Slept in. Went to the Counter for lunch and had a 1 pound burger,
topped with swiss cheese, grilled pineapple, dried cranberries, scallions,
roasted bell peppers, mayonnaise, and a fried egg. I finished it, but have no
intention of ordering a full-pound burger ever again.
I also learned enough about gstreamer to get audio playback of the Kinect audio
capture. Sending the data through ROS is still a work-in-progress; mostly because
packaging OpenKinect safely is funky.
Sunday: I rediscovered my Nokia earbuds in my backpack. I've been missing out
on proper bass response in my headphones all summer. I listened to a few of my
favorite OCRemix songs and got goosebumps from the warmth of the bass. I can't
wait until I can use my speakers back in Berkely again.
I was also semi-invited to a birthday party for Susie Fu, a TAMSter from two
years below me who went to MIT and is now interning with Microsoft in Mountain
View. She instructed the Brohaus to dress up, but all of my dress clothes were
in Berkeley, so I wound up wearing jeans, a button-up shirt, my red fedora, and
my trenchcoat. I had Chuan try on my trenchcoat, and the result was nothing
short of fantastic:
You want to trust this man with your money.
We drove up to SF for the house party, which was substantial in size. There
was a variety of food and drink, a number of people from TAMS, and lots of
people from Microsoft. I met another PhD student who works at MSR on theory
(in particular, asymmetric cryptography), and a pile of other PMs and SDETs. I
also ran into another guy from Lucasfilm that I'd met before at the Berkeley
recruitment event who had also been wearing a trenchcoat at the time. He
recognized me without my trenchcoat, and identified me as "the trenchcoat guy."
As a bonus, three other guys at the party were also wearing fedoras. Someone
else I talked to also knew Zach Abel. It's a small world. Most everyone was
pretty nerdy, and I enjoyed meeting them all.
Susie (or one of her housemates) has a very fluffy calico cat named Zoe. We
were fast friends. I miss having a cat.
Later in the evening, the partygoers played several rounds of Never Have I
Ever. Let's just say that there was quite a wide variety of things that people
had and had not done. Around 22:40, I called for everyone to help clean up the
place before we left, which was a great success. I washed ~2/3rds of the
dishes. I like being a helpful partygoer, and the hostesses were appreciative.
All in all, a great party, and I'm glad the rest of Brohaus talked me into
going even though I didn't have dress clothes.
Obviously, that won't keep the hopelessly addicted from their Internet fix. I
connected my N900 to my desktop and tethered. Then, I installed dhcpd and
bind9, configured both, and (after some iteration on firewall rules) got
network up and running for the other computers in the house. We're winners.
I made some good progress at work today on multiple projects. I am pleased
with myself, although one thing that we did is (necessarily) a pile of
ridiculous hacks upon other hacks. Still - they work (so far), so I'm not
going to knock it.
Chuan and I tuned the piano this weekend. It took 6 hours, and the top three
octaves are still somewhat flat (there's a particularly bad A and G up there),
but it's so much more playable than before.
We also went to Costco and bought lots of food. Apparently the pineapple
sausages I love so much come in packages of a dozen at Costco! :D
Sadly, the Costco we were at did not have the awesome chicken pot pies that I
recall from my time at home, but we did get some chicken alfredo pasta and
calzones, so those were tasty. Also: a giant container of strawberries. Think
4x what you see in normal grocery stores.
Yesterday was Katherine Maslyn's birthday. She and her housemate Amelia Lin
invited Brohaus over to Girlhouse for board games. Much fun was had - we
(Robby, Chuan, Dustin, Jeffrey, Arun, and myself) listened to Simon and
Garfunkel's best hits album on a real working turntable, looked at Dustin's
ridiculous Starcraft2 team page, went into Mountain View for bubble tea (and
ran into Kat Busch while we were there), played a new board game called 7
Wonders (Amelia won), ate cake, and generally just sat around chatting for a
while about all manner of things until 1am.
Tonight, we plan to go to a meetup on rationality and human biases organized by
followers of Less Wrong, a blog written by the same
person who wrote of Harry Potter and the Methods of
Rationality,
which is both absolutely hilarious and wildly informative. Hopefully the
meetup will be similar.
Tonight, Brohaus went out to dinner to celebrate the awesome summer that we all
had together. The full-time Brohaus residents took the interns out to Sushi
Tomi in Mountain View and ordered Omakase. As before, it was
fantastic, and we all had a wide variety of tasty dishes.
Chuan, Dustin, Jeffrey, Arun, and JJ
Jason, Brad, Michael, and Robby
Robby, Michael, and me
And again.
Bros chilling.
And bros laughing.
Some of the bros are addicted to bubble tea (particularly from Tea Era in MV),
so we stopped by there. On the walk back, we ran into Kanjun (who we'd met at
Susie's birthday party) and Raghu (who works for Facebook) on Castro Street.
We all chatted for a good ten minutes and laughed about how awesome our housing
situation was with 9 people in a house.
Then Robby and I went back to the office, since we'd only worked 6 hours that
day. I resolved a longstanding issue and documented my work.
Yesterday was my last day at imo. I wrapped up as much as I could, documented
the weird bits, and explained everything else to Ahmed. I really enjoyed my
summer working there.
Then I went back to Brohaus, put all my stuff (except my bed) in my car, drove
back to Berkeley, unpacked, and crashed on the couch. I had a hard time
getting to sleep; I went to bed at 1:30, which is a couple hours earlier than I
had been going to bed the past week.
I had to get up in time to arrive on campus by 8:30 (so early!) for the
training seminar for new GSIs. It was good - we covered lots of relevant
logistics, but mostly focused on how to be a good inspiring TA and how to deal
with situations that are likely to arise.
After the seminar, I went to BiD lab where I discovered that one of my
labmates, Kim, was to leave tomorrow to go work on designy things at
Disneyworld. Andy, Pablo, Kim, and I all went to
Ici (warning, terrible site design that
requires Flash), a fancy ice cream shop, to send Kim off. They had some
unusual flavors - I quite enjoyed my cone of Earl Grey ice cream. Then we went
back to BiD and set up the tables and net and C-clamps in ping-pong formation
and played all matchups. The graph of wins was a consistent
DAG - Pablo beat Andy,
who beat me, and I beat Kim. But we all had lots of fun!
I walked back home and set up my desktop. For a while, I thought both of my
graphics cards had died or something, but then I discovered that I had merely
switched the DVI cables and was looking at the wrong input. Then, after
cleaning up my room a bit more, I drove back to Palo Alto around 1am, since I
didn't fancy sleeping on the couch two nights in a row.
I slept in. Dustin and Jeffrey and I arranged plans to go to shopping centers
to get Dustin fancy things for a wedding he's attending. All our schedules
were a little wacky, but we managed to refuel my car, get lunch from
the Ace of Sandwiches, and head
out to the Stanford Shopping Center. Dustin was seeking a watch, possibly a
dress shirt, and definitely a tie.
Our first stop was Macy's. Didn't like any of the ties nor shirts, and it
turns out the watches were all in the women's Macy's. We stopped at the
Express (they have nice shirts!), but it was expressly for women, so no dress
shirts. We got to the women's Macy's and found the watch Dustin liked...for
$40 more than he could get it for on Amazon.
Then we all went to the Hair International and got haircuts. I had a few
inches taken off the back to get rid of all the split ends - I haven't had a
haircut in over a year. It feels better now.
Having failed to acquire a nice shirt and a nice tie, we drove up to the
Hillsdale Shopping Center in San Mateo. Since we were thirsty, we got frozen
yogurt, then proceeded to the Express which was not only having a sale but also
had a Foursquare check-in deal. I wound up getting a sharp red dress shirt.
Jeffrey and I helped Dustin identify shirt colors (he's colorblind) and he
picked one and a shiny black tie.
We stopped by the Lego store and reminisced about our childhoods. I
particularly appreciated the Imperial A-Wing:
Note the Death Star in the adjacent display box.
Then we went to Macy's and Dustin got a couple more of the stretchy athletic
Nike shirts that he likes, and I got one in dark blue. The Macy's apparently
requires presentation of photo identification for all credit card transactions,
which is at the very least in violation of their Merchant
Agreements with Visa and
Mastercard. I intend to report them - while the policy is perhaps
well-intentioned, it cannot be considered an effective fraud deterrant, and it
raises unpleasant questions about privacy and information security. I think of
the same sort of thing when I hear about government policies - oftentimes a
well-intentioned policy winds up accomplishing the opposite of its intent, or
fails to accomplish its stated purpose while causing unintended consequences.
We should create rules based on what they will actually do, not what we want
them to do.
Finally, we went to a Thai place in Mountain View for dinner, and then went
home and I stuffed my bed in the backseat of my car and drove back to Berkeley
and unloaded things and processed more backlogged TODO items.
"Someone in Iran is performing an active MITM on Google with this cert right now (remove diginotar from CA root list): http://pastebin.com/ff7Yg663"
Not a good sign.
In other news, I'm none too pleased with Apple right now. It's nigh-impossible
to get a MacBook Pro to boot from anything but the internal hard drive or the
internal CD drive. Which is a shame, considering the EFI fanciness they drove
to market.
I wouldn't have minded, except for the fact that there's a
bug in the Linux kernel
preventing the kernel from talking to the CD drive on my
MacBookPro8,2 at all. It's been around since 2.6.33 and thus affects at least
the past two versions of all the Linux installers that I'd deem suitable to
use. Moreover, there's a (two-character)
patch that I
hope will get merged before Fedora 16 Beta, but (as it was developed in the
last week) is currently unavailable in any distro.
So I can't read the installation stuff off the CD drive. Ordinarily, this
wouldn't be a problem - I've got at least three other methods for doing a Linux
install that don't involve the CD drive. People who have used Linux on
netbooks are probably familiar with the approach of doing installs from USB
flash drives. On the MacBook Pro, this is a non-starter - the firmware
can't boot from USB.
Now I'm about at a point where I'd try to just format the partition and do the
install as though it were in a chroot and rsync the filesystem from one in a
VM, except Apple also saw fit to remove their ext2 driver from OSX entirely.
So unless I want to install on either FAT32 (which will have broken
permissions) or HFS+ (for which I'd have to disable journalling to get
read-write support), or UFS (which admittedly would probably work, but I've
never set it up before, and I'm not sure if OSX will format a partition like
that), I can't do that either.
The other technique that I'd employ (network installation) has problems of its
own. When possible, I like to use the netinstaller off a USB drive, since it's
easier than setting up DHCP and DNS and NAT and TFTP and pxelinux and a
bootmenu...but we already established that USB is not an option. In
particular, I seem to have found a particularly annoying bug in either firmware
or an OSX driver. When I connect an ethernet cable directly between my desktop
and the MBP, suddenly, the Mac starts hanging in a strange way. Based on the
symptoms, I'd guess that every syscall is blocking and never returning. I'm
unable to spawn new processes, read or write on the filesystem, read network
config, or in general, do anything useful. This made testing annoying.
Searches on the web also suggested differing levels of success in netbooting
Apple machines from anything other than an Apple netboot server.
My options remaining are thus:
See if netinstalling something like Debian works and if so, do that and then update things and ignore the broken CD drive (I'm skeptical as to whether this is doable at all, at this point)
Install Fedora 16 Alpha in a virtual machine, build a custom kernel in that vm with that two-character fix, master my own installer ISO, burn that to a CD, and install from that media (more feasible and much greater probability of success, but an awful lot of effort, and downloading all the packages will push us over our monthly bandwidth cap (also not my fault))
Notify #fedora-qa, wait until this is fixed in Fedora 16 Beta, then install that (high probability of success, but won't happen for about a month)
So far, I'm leaning toward the last option, but the lack of a reasonable window
manager on OSX is killing me.
I'm also learning C#, since we're getting a bunch of Kinects for CS260 to do
projects. The Kinect SDK for Windows does not support such things as
Processing or Python (well, maybe IronPython, but I am NOT bothering with
that), so I'm going to learn enough to support the class for their projects.
Which means doing all the projects in advance myself. Which really shouldn't
be too hard - I'm at least halfway through the Bubble Cursor assignment
already.
Man, I should have taken out a long put option on VDSI yesterday as soon as I
heard about the fraudulent SSL cert their subsidiary issued.
The stock fell nearly 10% between when I read about the fake cert and now.
Wow, last night I had the WORST sleep I've had in the past year. Woke up at
least every two hours, and closer to every 15 minutes toward the end. Now I
feel nauseated, but not feverish. What the heck, body. We're supposed to be a
team.
I also dreamt vividly that when I woke up my computer and monitor were missing
(presumably stolen). This did not help.
This weekend was the Microsoft Puzzle Hunt. Ordinarily, you'd have to be in
Seattle/Redmond to participate, but this year they were kind enough to run a
simulcast instance of the game in the Bay Area through the internet.
It was great. The hunt was train-themed. I'd never hunted with any of the
people on my team before, but they were all pretty sharp. We started out
ranked in the top three, but lost some steam over the course of the day. By
3am, it was just me, Ankur, and Nima puzzling, and we were all pretty tired.
By Sunday at 5pm, we were in 9th place out of 29 teams. I quite liked the
round 2 meta - perhaps I should get a copy of Ricochet
Robots. I also
particularly appreciated the "Whistle Stop" puzzles, which were designed for
teams to complete together as quickly as they could which usually involved
plays on words. The first couple were quite easy and made for a good morale
boost. I can't wait for the MIT Mystery Hunt.
In other news, today I learned that there is a protein involved in the
mammalian signalling pathway called Sonic hedgehog homolog.
Yes, it was named after this one.
Oh, and apparently we humans have also made
glow-in-the-dark
cats. Pretty cool technique - they use an analogue of HIV for cats called FIV
to deliver genetic material to eggs before fertilization.
Today I learned how to write GUIs in C# with WPF. Never before has it taken me
14 lines to draw a single arc on the screen. I am used to
one, maybe two.
I am not thrilled.
Last night, I went to the opening concert of Dream Theater's US tour in SF. It was
nothing less than spectacular.
We lined up outside the Warfield a bit early. There were a number of tables
set up where people were playing chess and speed chess (for money, apparently).
I watched a game, and the players weren't particularly good - they missed a
free rook several times. I chose not to play.
Now, the cover band (Trivium) was...not my
kind of band. I couldn't understand a word they were saying (shouting?) over
the excessive amplification of the drums. In fact, I couldn't hear the guitars
over the drums either. I should have brought earplugs for this section of the evening.
Once Dream Theater came out, though, everything was awesome. They opened with
"Bridges in the Sky"
and fed that directly into "These Walls." The crowd responded with a standing
ovation and thunderous cheering. LaBrie said "now THAT is the warmest welcome
we have EVER had in San-Fran-Cisco."
The setlist included (I can't say I remember the particular order):
Build Me Up, Break Me Down (featuring a 4 minute drum solo with Mike Mangini, which was pretty great)
Endless Sacrifice
The Silent Man
Beneath the Surface
The Great Debate (a really interesting song about stem-cell research, with opinions from all over the spectrum panned left or right by their political stance)
Forsaken
A bunch of songs from their new album that I didn't recognize because I hadn't heard it before the concert
Encore: Under a Glass Moon (with the crazy awesome guitar solo)
Interestingly, no songs with lyrics by Mike Portnoy, who left the band in
September 2010.
I sang at the top of my lungs and drummed along to every song I knew. It was
awesome and amazing and joyful, and I was thrilled to be there. Those guys are
incredibly skilled.
In a completely unrelated court case, I found out that judges (from the 7th
Circuit, at least) can have a sense of humor: from
ProCD v. Zeidenberg:
There is no box; there is only a stream of electrons, a collection of
information that includes data, an application program, instructions, many
limitations ("MegaPixel 3.14159 cannot be used with BytePusher 2.718"), and the
terms of sale.
I see what you did there.
I recently installed KDE on Windows. I now have
Kontact syncing with Google
Calendar, a screenshot
utility that doesn't
involve me using an on-screen-keyboard to hit printscreen (Apple is really
annoying me), a hex editor
that rocks, and a reasonable fixed-width text
editor. I'm impressed - everything I've
used has been rock-solid stable. This was not the case a couple years ago when
I last tried KDE on Windows. Thanks to all the devs for their efforts.
As I was leaving campus last night, I saw a deer in the bushes. It proceeded
to run into the street and nearly got hit by a pickup truck. Oh deer.
My chair just broke. Yes, the one I've had for five years since TAMS. The
piece on the right-hand side that holds the chairback up snapped. It's served
me well for quite some time now. Time to search for a steelcase think on
craigslist.
Thanks to xkcd, I learned about the tragedy of the
commons. It seems to me
the idea applies to our banking industry, where the marginal value of the
addition of a new banker (or quant, or speculator, etc.) is great for the
banker, but we've already surpassed the percentage of GDP at which point
finance ceases to contribute materially toward GDP. But since the benefits are
privatized and the costs are borne by the public, any rational actor will do
what's worse for society. You could probably make the same argument about any
similarly "saturated" industry.
I just finished reading The Big Short. I
have little reason to believe that any of the structural problems with Wall
Street described therein have been resolved. Outlook: cynical and gloomy.
Had dinner with Matt and the Brohaus at Tofu House last night. Korean BBQ is
delicious. We started some ridiculous anime that had only six episodes, and
somehow everyone lost interest by the middle of episode 2.
Arun brought out his 5x5x5 Rubik's cube, which I spent a good while solving
twice. Kmas studied flashcards for Starcraft 2 with Terran units and
buildings, then played through a couple single-player campaigns. Jeffrey set
up his Kawai weighted electric keyboard and I played some random songs on it.
Another Brohaus intern moved in at some point that night after arriving from
Scotland. I crashed on the futon some time around 5am.
I drove down to San Jose to pick up a Steelcase Think chair from a guy on
Craigslist for the surprisingly low price of $250 (they retail for at least
double that). I now have a snazzy new office chair.
Yesterday, I went to
a talk by Ralph Nader
about the problem with Big-Time College Sports. It was a fantastic
talk - a very cohesive tale story detailing how the athletic scholarship and
intercollegiate sports have damaged physical education and intramural sports
for the rest of us, how athletes' top priority is not education but sports, and
how all this brings our society down what he termed the "sensuality ladder."
Today, I went to a talk on game theory in which
Tim Roughgarden explained that the
"price of anarchy,"
or inefficiency in letting a group of players all act toward locally optimal
strategies rather than having a benevolent dictator provide a globally optimal
strategy, is low. It also turns out that games that have a Nash equillibrium
wind up with the same price of anarchy as games that are merely "no regret" (a
much weaker assumption). A fantastic amount of this talk went over my head,
and Tim carefully avoided making any claims as to political outcomes of this
research, but I find it interesting to (probably improperly) apply this to
politics and say that the cost of libertarianism is "low."
Also today, Steve Jobs died. I
had the good fortune to grow up with a father who loves technology and an Apple
computer in the house. I wonder how different my life would be now, had that
not been the case.
It is upsetting to think that I can only be great at a handful of things before
I die. Thank you for choosing those things you found important, Steve. You
built a revolution in a garage.
I have about a thousand things running through my head right now. But first,
let's catch up:
My parents and little sister flew in last weekend + Monday (Oct 8, 9, 10) to
visit. It was a ton of fun. The weekend in summary:
Saturday: I picked them up from SFO and then we drove into the City to visit
Pier 39 and Ghirardelli Square. Traffic was the most horrible thing I have
ever driven through in my entire life. We tried to route around the traffic
and got caught up in a slew of streets that were one-way in the direction we
didn't want to go. We wound up driving (VERY slowly) along the Embarcadero.
There were crowds of people, all over - my mom described it as looking like
Seoul, Korea. On the upside, we were getting quite the tour of SF, but we had
no idea why traffic was so terrible.
Then I saw the B-2 stealth bomber.
It turns out that this was the tail end of SF Fleet Week, and there was a wicked-cool
airshow. We found parking and got clam chowder in bread bowls from a street
vendor, and then enjoyed watching the Blue Angels and Canadian Snowbirds do
incredible formations. We walked about Pier 39, and I got to see F-16s do snap
rolls while flying straight upwards and diving straight down. It was downright
amazing. Tommy Kennedy would have been in heaven.
We walked over to Ghirardelli Square and my dad and sister bought chocolate
while my mom and I just sat on the steps for a bit. We left the City just
after 4pm. We had dinner reservations in Sausalito for 6pm. We called to
warn them we'd probably be late. Horrible, awful traffic. From my
friends' reports, BART was also very crowded.
While we did make it into Sausalito right at 6pm, we stopped at the scenic
viewing area at the north end of the Golden Gate bridge for some fresh air and
photo opportunities. Oh, I should point out that my dad loves taking photos
and home video.
Dinner at The Spinnaker was
delicious. We took the Richmond bridge back to Berkeley. I showed my parents
my apartment and we talked and played cards. I took them back to their hotel,
and they gave me a box of Crunchy Corn Bran cereal - apparently, my dad can
only find it in bulk in 24-packs nowadays.
Sunday, my family went to the nearby Trader Joes and obtained breakfast. I
picked them up from near their hotel and we ate at my apartment. We drove up
to the Berkeley Rose Gardens and spent a good while there. It was a lovely day
out - a little warm in the sun, but perfect in the shade.
I pointed out some impressive-looking spiderwebs. My dad went wild with his
optical recording devices, and my mom quipped that they'd get home and have a
half-hour of video showing nothing but spiders. It was fun, all the same.
We had lunch from Gregoire with
the company of a long-time family friend Anna Theiret. After lunch, I took my
family to campus and gave them the tour of the four buildings that I spend 98%
of my campus time in and the student center. The BiD lab was surprisingly full
of people for a weekend. Lora would later describe my family as adorable,
particularly my dad shooting home videos.
The family went to Crepevine for dinner, which was also great, then played
cards until late-o-clock. I haven't played very many 4-player card games
lately, which is such a shame. Gotta get Andy and a couple other folks to play
bridge. Alternately, we should get our D&D campaign back up and running,
once Wes does his quals talk.
I got up early to visit my family at their hotel one last time before I had to
leave to get my things and go to class. They successfully navigated BART to
the SFO airport.
The week began.
I submitted a (rather short) proposal for a talk for 28c3 on reverse
engineering USB protocols. If accepted, then I have one less thing to worry
about (getting a ticket to 28c3 before they sell out) and one more thing to
worry about (fleshing out said talk before the end of the year).
I have been working extensively on a project for the UIST 2011 Student
Innovation Contest which I will disclose in more detail after we present it.
Those of you who stalk me on github may have noticed the repo with my code
already. I'll see if I can take a video of people using the contraption and
post it here.
I've been updating my collection of OCRemix
tracks and discovered that a bunch of folks whose remixes I've long loved now
have their own professional discography. Many of the people that started out
reimagining game music for the love are now employed by the game industry as
composers.
Andrew and
Jillian Aversa (OCR:
zircon and
pixietricks) do some
of my favorite electronica and celestial vocals, respectively. I bought all of
zircon's discography; from the first two sets alone, I can conclude that I will
be enjoying these for a long time to come.
I noticed at this week's BBQ that one of the Berkeley CS grads has sectoral
heterochromia - in
this case, a brown spot in an otherwise blue eye. I found it fascinating, so I
looked up a bunch of reference on heterochromia and eye genetics. After doing
the requisite reading, I wonder if I might have central heterochromia.
Apparently, it's pretty common for people with green eyes. What do you think:
do I have heterochromia?
I (finally (this has been a work-in-progress since late April (let's see just
how deep I can nest parentheses (this looks kinda like Lisp)))) merged
support
for image registration
into libfreenect this week. Special thanks go out to
Kyle McDonald and
Florian Echtler for their
contributions to this (somewhat large) patchset.
I both hosed and recovered my laptop's Windows install today. Don't install
self-signed USB filter drivers in 64-bit mode if you haven't turned off signed
driver enforcement. Particularly if both your mouse and keyboard are USB and
your machine (MacBook Pro!) has no legacy ports. Safe mode won't save you, but
the advanced boot options menu will - it has "Disable Driver Signature
Enforcement" as one of the boot options. Huzzah, all works. It's worth noting
that none of the tools that use event tracing for Windows (ETW) can actually
show you the data in the packets that your machine sends, since the kernel
doesn't log that data, but filter drivers (like BusDog) can.
I went to the local Thai restaurant for dinner tonight. The waitress recognized
me and knew my order. I had my food in something like 6 minutes from when I
entered the restaurant. I think this is the first restaurant with which I've
developed that kind of rapport.
Tomorrow morning I take Amtrak to Santa Barbara for UIST. I plan to write a
Wikipedia article and grade assignments en route.
Got back from UIST last night. Storytime! (Note: post-dated. I'm really bad
at getting these entries written on time.)
Sunday: What a terrible morning. I managed to hit my snooze button 9 times
unconsciously before finally waking up 70 minutes later than planned. I was
supposed to wake up at 7:15, so I'd catch the 8:08 BART to Oakland and catch
the 8:50 Coast Starlight Amtrak train (which leaves once per day).
I looked at my alarm clock, and the last digit was right. Unfortunately, the
other two were too large. I had missed the BART. Next train left Berkeley at
8:38, and arrived in Oakland at 8:51 - too late. ****.
Turns out I wake up very very quickly and become quite alert when I'm dosed
with adrenaline like that. It was a good thing I had packed my bags the night
before - I threw on clothes, looked up driving directions to Jack London Square
on Google Maps, and snapped a picture of the directions with my phone. Google
estimates the route at 17 minutes. Sounds doable. I drove as quickly as
safely possible to my destination.
I arrived by the station, but didn't know where I was supposed to park. There
were garage buildings, but they all appeared to require permits. Driving
around frantically, I finally saw a sign that read "Amtrak Parking" and
followed it. As I crossed the tracks, I saw the train I'm supposed to be boarding
waiting. I got quite nervous as I had to drive further away from it.
I turned into the nearest parking lot...which happened to be owned by Fox News,
and gated. I exited that parking lot and entered the next one, which had a
pre-pay meter where you buy a permit for X hours and display the receipt on
your dash. Great, except that it only lets you pay for 72 hours in advance. I
was going to be gone for ~86. I decided I would worry about that later, and
right now I needed to catch my train. I bought the 72-hour permit, unloaded my
backpack and suitcase, and sprinted as fast as I could.
I all-out ran a football field and a half to the four-story skybridge that
would take me back over the tracks. I reached the base right as the elevator
started going up. Guess I'll take the stairs. I was breathing heavily by this
point, but I couldn't stop - I had to catch that train. I took the stairs back
down on the other side, and ran for the end of the train where I saw a
uniformed someone standing. He asked if I had a sleeper car ticket. Nope.
Wrong end of the train.
Which was fine, because I didn't have my ticket anyway. The woman at the other
end of the train informed me that 1) I needed a ticket, not a confirmation
code, so I'd need to go inside and print one off, and 2) She had to leave. So
I left my bags right there, ran in, printed my ticket, and ran back. I
collapsed into an empty seat, lungs burning, asthma attack setting in. Within
30 seconds, the train started moving. I was exhausted. But I was on board.
The rest of the train ride was fairly uneventful. I used my N900 to set up a
wireless hotspot for myself and Valkyrie, and I did some reading. Lunch was
expensive and the burger was a bit dry, but I've had worse.
We arrived so early that the previous Amtrak train was blocking us at the
station in Santa Barbara. I checked into my hotel and dropped off my bags, and
we walked the ~2km along the beach to the conference hotel to collect our
registration materials. All the food was gone, so we went to El Torito for
dinner. I told everyone about the origins of the
CD-ROM boot standard
by the same name.
Monday: overslept again, but it just so happened that the session I was late to
was the crowdsourcing session, so I wasn't too heartbroken. UIST is a
single-track conference, which threw me off, initially - there's not a whole
lot going on at any given time other than the paper presentations or breaks.
On the upside, you get to see all the papers. On the downside, the long
sessions get somewhat grueling.
I particularly liked Leila Takayama's talk about the (in)effectiveness of
sidetone for telepresence and the MIT folks' Real-Time Collaborative Coding in
a Web IDE talk. I want better tools for collaborative reverse engineering.
Björn,
Wes,
Meredith Morris from MSR,
Thorsten and
Jan from Aachen, Valkryie,
and I all went to the nearest restaurant for lunch (which happened to be that
same El Torito). It was a pleasant discussion - I really enjoyed talking with
Thorsten about differences between the US and Germany/the rest of Europe.
Monday's keynote by Ge Wang on his iPhone Ocarina and ChucK and thinking about time-based
concurrent programming was fascinating. I thought back to the time when I was
chatting with Dave Thomas about it - he found it neat and was tinkering with
randomly generated melodies. I wound up using ChucK to make TDMF tones - at
some point, I'd like to be able to identify phone numbers by ear.
We presented our project - the Moussage mat - at the Student Innovation
Competition. We take input from a Microsoft TouchMouse, transmit it over the
network, map the touch points into touch intensity within 12 zones, and send 12
values to a microcontroller, which drives 12 vibrating motors embedded in a
sheet of foam that you're lying on for a human-controlled massage experience.
It was very well-received - we wound up with a bunch of people claiming they
would just lie down and sleep there for a while. (Nice try.) By the end of the
night, people were coming over, saying that their peers had told them to come
see our project. People were amused that we had used bullet vibrators as
motors. They're cheap (the ones we got were ~$4 each), small,
well-encapsulated, and reasonably powerful. In other words: the perfect
off-the-shelf part for our project.
Sadly, we managed to burn out the power supply that we were using to drive all
the vibrators about 70 minutes into the three-hour demosession, but that was
after the judges (and most of the people that would visit our demo) had come
through, so it turned out okay.
We're in the process of preparing a demo video; I'll probably make a decent
writeup of the whole thing and post it as its own subfolder on my site.
Later.
Tuesday: More fine talks. I liked the Stacksplorer talk.
Julia Schwarz likened dealing with
multitouch APIs as trying to do a lot of things at once, like juggling - and
then proceeded to juggle 5 balls on stage. The round of applause was merited.
We talked after the session about other approaches people have taken and how
all "potential action" previews need to be free of side effects. Other cool
talks:
PocketTouch: Through-Fabric Capacitive Touch Input. I'd love to be able to
manipulate my phone in my pocket. I already don't look at my keyboard when
typing, why should I have to get my phone out to interact with it?
Conte: Multimodal Input Inspired by an Artist's Crayon. I was impressed by
just how fluid the user in the demo video was with this tool. It gave me hope
that some day, these technologies will enable a new level of masters in their craft.
Calibration Games: Making Calibration Tasks Enjoyable by Adding Motivating
Game Elements. The speaker was wildly entertaining.
After the talks and a brief town hall, it was dinner time. I had several
helpings of the delicious salmon. The keynote speaker, Dan Jurafsky, gave a
great talk that explored language and what we can learn from its history. We
traced through the history of "ceviche" and "ketchup", explored the difference
in words used to advertise cheap and expensive potato chips, and learned some
decent predictors of interactions from a corpus collected from a speed-dating
session. I will be hopelessly analyzing all my conversations henceforth.
After dinner came thanks and awards. Our Moussage mat won 2nd in the "Best
Implementation" category of the Student Innovation Contest! I was thrilled.
After the level of positive reception from the previous night, I somewhat
expected we'd win something, but it was exciting anyway. I wound up chatting
for a good while with Juan, the guy from København whose project beat
ours. Maybe I'll see him during my Europe trip this winter.
Since it was our labmate Shiry's 32nd birthday, Bjoern invited us back to his
hotel room for birthday festivities - the party included Bjoern, Shiry,
Maneesh, Wes, Peggy, Valkryie, Nick, Manas, and myself, all from the BiD lab.
Manas brought along Jen Fernquist.
Fun discussion was had - politics, economics, and all sorts of things I usually
don't chat about with my labmates.
Manas, Jen, and I went to the Student Volunteer party after that. Katrina, the
head SV who I'd met in May at CHI, had told me to come and that I should feel
free to "bring all the cool people." Most declined. I volunteered to be the
designated driver - I've never been much of a fan of alcohol. Others at the
party were.
Wednesday: I was originally going to take the Amtrak back on Wednesday, but
then I found out that the 3D interaction session wouldn't happen until after
the train left. That session included the much-anticipated KinectFusion demo.
Fortunately, Manas had driven down and was kind enough to offer a ride in his
car back. So I got to see all the fancy demos!
Since relatively few people get to go to these conferences, let me say very
clearly: KinectFusion works just as well in person as it appears to in the
video - 30 FPS with all the bells and whistles. It crashed a couple times, but
I'm not going to complain about it. Very impressive work.
I liked the concept from 1 Thumb, 4 Buttons, 20 Words Per Minute: Design and
Evaluation of H4-Writer of making Huffman trees with four branches at each
node. It makes me want to learn all the DTFM tones by ear and use them as a
higher-data-density Morse code.
The Toucheo
demo was just awesome. Multitouch + 3D display, with somewhat sane surface
manipulations. If they throw in a Kinect for direct 3D interaction, they'll be
golden. I loved it.
Manas and I drove back to Berkeley, stopping in San Luis Obispo for dinner. We
found a place called Thai Boat with unusually enthusiastic staff. They were
all just so happy to serve us! And you can never go wrong with pineapple fried
rice. We stopped at a gas station, and the owner of the convenience store
greeted Manas in Hindi and they had a nice conversation. Shared language and
culture gives most people a sense of connectedness, it seems.
We got back to Oakland shortly after midnight. We found my car, which to my
great surprise was neither ticketed, towed, nor impounded. I drove it back to
my apartment in Berkeley, relieved, and crashed.
Over the past couple years, I've come to an interesting realization - I like
knowing how protocols and markup languages work. Their design informs my own,
and being able to speak the exact language that another system is expecting
means that you are able to interact with it. You are able to understand it.
You are able to analyze its behavior. This concept of common ground - speaking
the same language, whether it be text strings, or a sequence of voltage levels,
or something else - is a very useful abstraction. It makes it possible to
build higher order things on top of them, and then people only have to
understand the layer they're working on.
One of the reasons I think I like knowing all sorts of protocols is that it
means that I can interact competently with systems (or people) at any of the
layers. It means that I can be useful in a greater variety of settings. It
means that I can start to identify where in the stack of layers a brick is
missing, when something breaks. And that is something the world seems to find
valuable in an engineer. If I were a recruiter, I would expect the breadth of
an applicant's competencies with different protocols and languages,
particularly in different layers, to correlate with the applicant's ability to
deal with complexity, which I'd consider fundamental to good engineering.
It would be terribly convenient if we could apply the same principle to people.
And to some extent, we can. Some layers, like spoken and written languages,
are fairly good opaque abstractions. Others, like cognitive models, emotion,
group interactions, and a pile of the other interesting emergent behaviors of
humanity are not. People can't handle that sort of cognitive load without
better abstraction - we are simply too complex.
But since we have to deal with other people anyway, I think our brains come up
with ways to cope. Inevitably, we wind up making simplifications to systems.
I'd posit that both stereotypes and common spoken and written language are a
way that the human mind deals with the excessive complexity of people. Through
both, we have restricted the amount of complexity that an individual must
consider when interacting with people.
Both have an up-front cost to learn and a loss of a certain amount of
expression as a result. People have feelings that they cannot adequately put
into words, and stereotypes can result in improper and hasty judgment. But in
exchange, we gain an abstraction layer - a common ground - that makes a great
many other interactions wildly more efficient or even tractable. Without
written and spoken word, we would lose history. Without history, we would be
unable to build on the knowledge (and abstractions) of our predecessors.
One last thing that we rarely mention: once fixed, abstractions cannot be
easily changed, since other abstractions are built on top of them. Just as we
cannot change the USB protocol (we can merely add a new one), we cannot easily
change languages, or alphabets, or stereotypes, but we can add new ones. Just
as the tail and the appendix were useful in our earlier evolution, we are stuck
with these eccentricities in our development. We must design our abstractions
wisely, if we can, for we are stuck with them once we make use of them. My
mind is filled with questions:
What other abstractions exist?
Can building more or better abstraction layers for how people behave improve our interactions?
Does building these abstraction layers come at too high a cost?
What new abstractions are still in their infancy?
How can we shape them to make them more efficient, to give them lower costs?
Can any of this actually be reasoned about from our current layer of abstraction?
Awesome party last night, hosted by the BiD lab. We arrived early and
rearranged the lab to block off our desks and have more open space in the
middle of the rooms, and set up lights and drinks and food. My advisor Björn
was our DJ. I should point out that Björn
was a professional DJ before he went to grad
school, and it was the interesting interfaces used by DJs that prompted him to
do research in Human-Computer Interaction. Needless to say, the music was great.
Seeing Björn in the mix, I couldn't help but fondly recall Gabe's random mixes
on hall at 5E, and learning the general ideas behind DJing - beatmatching,
pitch control, lead in, cueing, and crossfading. Björn says that most people
take about a year to develop the tacit knowledge needed to be able to mix well
and consistently. I wonder if that'd be a skill worth investing some time in.
I'd guess that another large part of it is getting to know what BPM and key
each of your songs are in. Time to go install mixxx.
As this is the weekend nearest to Halloween, the party was a costume party.
There were some great costumes - Manas printed off a poster-sized blank
Facebook Wall using the lab plotter and let people write on his wall. There was a
girl dressed as Siri. There was a vampire (and not the sparkly variety) and
several pixies. Luke was a crossing guard. Pablo was a convincing Don Vito Corleone,
with a coat and white vest. I wore a red dress shirt and dark dress pants, my
red fedora, and my black trenchcoat. I entered the costume competition for
"Geekiest costume." People seemed confused by my costume until I explained "I'm
here representing Red Hat."
Then I got thunderous applause and won. About five minutes later, someone
walked in dressed as a Rubik's cube with faces that lit up and changed colors,
and I passed my prize on to him in deference.
A number of people thought that I was trying to dress up as Carmen Sandiego, since
she wore a red fedora and a trenchcoat (though hers was red, and mine was
black). In a curious coincidence, Sarahdid dress up as Carmen Sandiego. Peas in a pod, I tell ya.
A good-sized crowd came to the party - around 30 or so people at max headcount,
though more than that cycled through over the course of the night. We made
sangria with lychees stuffed with blueberries, intended to look like eyeballs.
The whole event was described by Andrew as "the best departmental party I've
ever been to." Pretty awesome. Thanks to the CSGSA for sponsoring, and everyone
who came for attending!
I want to carve a pumpkin into a Guy Fawkes face for November 5th. I might
actually do this - I think BiD has some leftover pumpkins.
In other news: I went through my inbox and reduced 2600 messages to 79 (most of
which I still need to follow up on).
In more other news: I need to do a writeup on this website of the Moussage
project. And I need to do a writeup on
libtouchmouse, which I've produced
over the course of the past week.
Friggin' finally: 28C3 tickets preorder
information. Still can't buy a
ticket until next Sunday, but believe you me, I know where I'll be at 1pm PST
on 2011-11-06.
Second: happy birthday, Kinect! (Incidentally, it was released one year ago,
today.)
Today was a somewhat busy day. I moved the contents of my savings account from
BoA to a credit union (because Bank of America
stealshouses). This caused trouble on both ends.
At BoA there were some complexities:
I told the teller how much I wanted to withdraw. She replied: "Uh, we don't have that much in cash." A cashier's check will suffice, thanks.
I opened the account in Texas, but was drawing funds from California. Since there are different regulations in each state, they don't have as much access to the account information in CA as they do in TX.
In the 5 years I'd owned the account, I'd never once written a check, so the teller couldn't find my signature on file online to compare against the one I was providing to authorize the withdrawl. Eventually she wrote something on the withdrawl form, and proceeded with the transaction.
Somewhat surprisingly, I didn't have to provide any form of ID (although I
suppose debit card swipe + PIN counts).
Making the deposit at CUBS was comparatively simpler, but took just as long -
the teller had to get an override from her manager or another branch to be able
to deposit a check that size. None of the other branches were answering the
telephone, but after about 20 minutes her manager came in and took care of the
deal.
I met with Dave Wagner today, and over the course of two hours, we planned out
a research project, came up with some interesting questions, talked to several
other students, realized that the questions were not actually interesting, and
came up with some other questions. The preliminary results from one of the
other students' studies make me disappointed in cell phone users at large, but
even more so, at Google for failing to let Android users make more granular
security decisions. Permissions should be something you can dynamically grant
or revoke at runtime, and app developers shouldn't be taught that they can
write code that assumes it can get whatever privileged data it wants at
runtime. Blargh.
After that meeting, I met with Ankur, Annelise, and Josiah to plan puzzle
hunts. We'll be rerunning the 2011 Berkeley Mystery Hunt this winter for the
general public, so there was some discussion of logistics and review of the
puzzles to make them more suitable for the anticipated audience. We're also
laying the groundwork for what will hopefully become the 2012 Berkeley Mystery
Hunt, which is super exciting because one of my life goals is to write and run
a puzzle hunt. We have some great ideas, and it's really fun to brainstorm
with people who are just as excited about puzzles and twice as skilled. We had
three solid hours of creative discourse, and my mind is still brimming with
other crazy ideas to try. I love this feeling.
The first batch of 28c3 tickets sold out 3 minutes after being made
available. 800 tickets.
180 seconds. All I was able to do was fail at refreshing the page since their
servers got seriously
DoSed by all the
would-be Congress-goers. For the first two minutes, I couldn't even open a
socket to the presale website, and then I got a bunch of 502 Bad Gateway errors
after that. When the system finally managed to reply to me, it was clear that
there were no tickets available. I was sad.
I went to lunch at a sushi restaurant in downtown Berkeley. They had this
absolutely amazing baked salmon roll - it was baked salmon with some sort of
mayo-based sauce and masago atop a roll containing crab, cream cheese, avocado
and asparagus. It was one of the most delicious things I've eaten in recent
times.
When I got back to my apartment, I figured I'd seek more information about what
the heck had happened with the 28c3 tickets from more official sources (when I
left, their blog was also down, probably due to DoS). The comments confirmed
that the batch of tickets was indeed sold out. Then I logged into the 28c3
talk submission page, just for kicks. I hadn't received any email from the
organizers, and I'd read that speakers had already been contacted.
But then I saw this:
And I clicked on it, and saw this:
So, uh, I guess I'm going after all, and I'd better start preparing that talk
soon!
Not going off Daylight Savings time. Not when the stores start decorating for
the next batch of holidays.
I'm referring to the couple months during which every Wikipedia page is graced
with the portrait of Jimbo Wales as the nonprofit Wikimedia foundation runs
their annual fundraiser. It's the time during which I'm constantly reminded
that I stand on the shoulders of giants, and that which we may think of as
"free" is only so available thanks do a great deal of labor and love.
The fact that I assume that Wikipedia will always be there at the tip of my
fingers is nothing short of breathtaking. It's funny that I can get far more
value out of Wikipedia than I do from the majority of my textbooks. I can read
about practically anything that I want to know about. And in the rare case
where I find something that I think I could explain better than this wonder of
the world already does, I can share my knowledge with the world. What a
beautiful thing!
I can't imagine a world without Wikipedia any more, and I hope my donation and
yours can make it so that we won't have to.
Yesterday I got to help Manas break into his own apartment after he locked his
keys inside and himself outside. Fun times!
I went to the local Thai place. I ordered my food at 21:26. My food arrived
as my watch ticked 21:30. I was impressed.
Oh, look what we have here. A proposed
constitutional amendment to get money out of Congress. Say, weren't those
Occupy protesters calling for something like that?
Thanksgiving was a blast. After thoroughly wrecking my sleep schedule after
the systems take-home midterm (I wound up falling asleep at 20:30 and waking up
at 01:30. Oops.) I managed to get back to sleep around the sunrise at 06:00
and slept until something like noon. I baked 8 miniloaves of bread and
pineapple bake (recipe!) for a potluck gathering
of friends.
There were about 35 RSVPs, and there was food aplenty: three turkeys (two fried
in oil, one roasted in the oven), three different kinds of mashed potatoes, two
kinds of green beans with cheese, sweet potatoes, sweet potato falafel, several
salads, and at least three pumpkin pies.
Yum.
This was just the main room; lots of people were in the adjoining rooms.
I had the pleasure of meeting half of the AeroFSteam (I met Yuri, Mark, and Allen).
I'll get to meet the rest of them on Tuesday when I visit their office in Palo Alto.
Andrew and I hung around until late-o-clock, eating way too much food and
chatting/singing/playing guitars with folks. David Moore brought out his
fiddle, and he and Shaddi were playing folk songs from the South. Good times.
MBR: Submit a 512-byte MBR boot sector that does something interesting (the
specifics are up to you) and before exiting prints "Stripe".
This seemed like a good choice because I'd somehow managed to avoid learning
x86 assembly up until this point - I can deal with ARM and MIPS and Zilog z80,
but I never learned how x86 works. (And boy, is it funky.) Anyhow, knowing x86
assembly could be helpful if I ever want to reverse-engineer object code for that
platform.
Someone I know from A&M who goes by the handle Pathore helped me understand
the basics of x86 assembly, BIOS calls, string handling routines, and helped me
debug code over IRC. For several hours. It was a very intellectually stimulating
experience. Pathore, if you're reading this, I owe you dinner or something next
time I see you.
Once I'd picked up enough of the basics to be dangerous, I wound up making a
stupidly simple game where you move a character ("%") around a grid collecting
the letters of the string "STRIPE". It fits (exactly) in the 440-byte code
segment allowed in the 512-byte MBR, though if I really wanted I could shorten
the printed strings and add more code. It's quite heavily optimized for size,
and I'm fairly proud of a couple of the things I thought of to save bytes.
It's not something you usually have to think about any more these days.
My submission is here,
for those who would find its workings interesting. It's heavily commented to
help you (and me) through. Let me know what you think!
Today, while walking home from campus, I saw not one, not two, not even three,
but four raccoons. The first two were poking about the garbage cans
next to a house which had a sounding burgler alarm. I suspect this was the
doing of the coons.
The second two were in the yard of a house outside which I often run into a
friendly cat. I heard a rustling sound, pulled out my light, and sure enough,
two more of those faces staring at the bright thing. Crazy.
I could tell it was that time of the year because there were some 20 people in
the BiD Lab today. Yes, on a Sunday. I'm personally trying to finish up:
An interesting distributed systems project
IP law assorted things
Giving good feedback to the 15 teams in the HCI class I'm TAing (and providing whatever other resources they need to succeed)
Finding a job at a top-notch startup, which implies learning more about stock, options, and trying to compare benefits. Also interviewing. Also applying.
Booking plans for my Europe trip in less than a month
Preparing a quality talk for 28c3
Oh, research towards my masters project is probably supposed to be on this list somewhere.
So, uh, that's why I haven't been updating as often as I'd like.
Zeus, let all this madness be over. My sleep schedule is SO hosed right now.
Friday night/Saturday morning was a total lunar eclipse. I woke up at 04:00
and watched part of it from my bed and part of it from my balcony. I tried
taking pictures - since I didn't have access to a telescope like I did
last year, I instead tried setting my point-and-shoot
on stable surfaces and doing long exposures. The results:
The house across the street.
Me watching from the warmth of my bed. This was a 15-second exposure, and my room is actually illuminated by my monitor with the backlight on but all pixels black. I had another picture that was with some of the pixels white, and it looked like daylight. Long exposures are crazy!
The moon itself.
There wasn't really time to sleep much after the eclipse, since Saturday was
the rerun of the Spring 2011 Berkeley Mystery Hunt (for a wider Bay Area
audience). I packed my dual laptops up and set up shop in our headquarters.
Ankur arrived, and then Josiah and Annelise and I fetched a bunch of food
(muffins, croissants, bagels, clementines, milk/juice) from Josiah's cubicle
for all the attendees.
Kickoff was a little worrisome - we had more interested teams than available
rooms to give out. It's troublesome that student groups aren't allowed to
reserve rooms on campus during R&R week.
The hunt itself started off slowly, but got more hectic as time went on.
Some of the submissions were hilarious, particularly for the Mario Savio puzzle.
When it came time to set up the runaround, things got a little frantic - Ankur
went to the start of the runaround to meet the team there and follow them
around on the runaround, while Annelise watched from one stop ahead and Josiah
ran ahead deploying envelopes at each of the stops. This was fine until a
second team reached the runaround and Annelise had to go join them, at which
point we had no more margin of safety, should a third team reach the runaround.
Since we only had three people left at HQ (myself as coordinator and
email-answerer and two people running the answer queue), I became nervous as a
third team completed every remaining puzzle in the game. Fortunately, half of their
team was driving in from San Francisco, so Josiah was able to finish the deployments
in time for us to make it back to the runaround start before the team arrived.
One of the teams figured out the final puzzle minutes before the hunt ended at
10PM, which was perfect. All in all, the hunt ran fairly smoothly, and I had a
lot of fun seeing things from the other side. It was great hearing from the puzzlers
what they liked and wished was different about the hunt, as well as advice from seasoned
puzzle-writers on how to plan such things. Good times.
The hunt itself is archived here, if you want to play along. My favorite puzzle was
Cipher. :)
Soon, the Campus League of Puzzlers will flesh out our plans for the Spring
2012 hunt. I am excited!
So that was Saturday. Sunday consisted of going into lab and tooling away for
the entire day at our CS262 project - a flexible distributed architecture for
computing similarity matrices with fast partial availability of results - with
Luke and Pablo. An all-nighter was involved. We produced useful results some
time around 7AM on Monday. We finished our poster at 9AM. The poster session
started at noon, so we each went home and slept for something like one hour,
changed clothes, and came back.
Somewhat surprisingly, the poster session was a
bring-your-own-posterboard-and-stand affair. Our presentation to Eric Brewer
was something like third-from last, and didn't happen until 14:30 (at which
point we'd been manning our poster for two and a half hours). On the upside,
that meant we'd gotten a lot of practice at explaining our system. It was
well-received by both faculty and grad students alike, and a lot of people had
useful ideas to improve our system.
After the session, I went with Luke, Pablo, Shiry, and Shaddi to Barney's
Gourmet Hamburgers (which are amazing, for the record). We decided on the
location by using the Paxos protocols
to reach consensus. If you know what Paxos is (and how complex
it is, and the implications of using it in the real world between people), you
should be laughing right now. Because we sure were.
After dinner, I went home and slept. My schedule now aligns more closely with
those of people living in England. I woke up around 02:00, and then started my
take-home IP law final, which was due at 9AM. There was much quoting of the US
Code and explaining why various scenarios would probably turn out in particular
ways.
That was this morning. I'm grading lots of final papers for CS260. I am very
ready for this to all be over.
I successfully completed my slides and gave my 28c3 talk titled "Reverse
Engineering USB Devices" today. Saal 2 was completely full. I'm going to
guess that the room held between 400 and 500 people. The talk went better than
expected! bushing and pytey (the lead devs on the OpenVizsla USB sniffer project) came up and
introduced themselves after the talk. Another guy gave me a ProgSkeet board, which will be neat to play
around with. I imagine this will come in handy when the Wii U comes out. :)
There was a small crowd - 10 or so folks - who came and chatted with me at
greater length after the talk. One of them gave me a USB controlled battery
charger, accompanied with the story of its existence - apparently the only
release of the software that lets you see how far the batteries have charged
includes some spyware. In addition, Energizer "resolved" the issue by simply
taking down the download from their website. Ha!
Another couple guys told me about their project,
netzob, which sounds promising. I'll have to
investigate in greater detail, but at a first glance, it seems it implements a
lot of the things that I called for when I gave my vision for what better
reverse-engineering tools would look like. Awesome.
One last guy chatted with me for quite some time. Turns out he wrote the
"forcedeth" Linux Ethernet driver for nForce chipsets.
I spent a good portion of the rest of the day hanging out with the folks at the
fail0verflow table. marcan and another guy were tuning a laser projector
powered by
openlase.
bunnie was working on his awesome
NeTV. pytey told me about the issues the OpenVizsla boards had had, and gave
an update on that timeline. bushing was busily stepping through format string
exploits. Such fun!
Wow. The past four days have been a nonstop high. What a wonderful way to end a calendar year!
The third and fourth days of the CCC were as fantastic as the first two. I met
with the lead developers of sigrok and
netzob, and introduced them to each
other. I finally saw a Beagle 480 in real life (that thing costs
1400USD...thanks, pytey!) I learned more about how NAND and NOR work. I have
a lot of talks to catch up on. I did make it to
Harald Welte's talk
about cellular protocol stacks for the Internet, in which I
learned about 60 acronyms in 60 minutes, and that I should start reading all
the thousands of documents pertaining to cellular network protocols. I also
got to hear Bunnie Huang's talk describing a man-in-the-middle attack on
HDCP,
and his completely legal, noninfringing, legitimate business application of the
HDCP master key. Since conferences are about people, not just talks, I spent
the majority of my time outside the presentation rooms, chatting with folks,
thinking about open problems, or hacking on stuff together. I believe the
general policy I heard from Andy Carle regarding CHI is "if you're going to
more than two talks per day, you're doing it wrong." Note to self for next year: don't plan
to hang around Europe for 11 days after the Congress again; you need that packing space
for electronics, not clothing.
After the Congress ended, we packed up our things and a decent group of us (6)
went out for sushi for dinner. The sushi was mediocre, but the fellowship was
great. After dinner, we walked to the
c-base which was configured as a bar/dance
club for the evening. The music reminded me of the parties on 5E. I had a
couple of the traditional CCC drink, called a
tschunk. I chatted for a good
while with many of the folks I'd met, and we exchanged (or updated) contact information.
The c-base also had a multitouch table, on which people collaborated at a game
of Untangle, or played 4-player Tron, or multitouch Pong. Reminds me, we
should get the BiD multitouch table working again.
As it is the tradition of the console hackers to do something awesome to the
bar computer each year, marcan and I took over the bar computer and got to
work. The machine in question is this ancient Via C3 Samuel 2 machine (it
doesn't even support cmov!) packed inside a C64 case, hooked up to a CRT
monitor. The keyboard was replaced with a USB keyboard taken from an iMac G3.
It was running some slooooow version of Puppy Linux copied into RAM from a 1GB
USB flash drive. It didn't even have python. Nor did it have a functional
package manager.
First, we installed Debian on marcan's 8GB flash drive. This was nontrivial.
marcan wound up fetching the cdebootstrap .deb (and several dependencies),
hexdumping the file and searching for the gzip header for the tar file
contained therein (because dpkg didn't work) to do an incredibly hackish
install of cdebootstrap. From there, we had to alias gpg to /bin/true, because
gpg was using a libcrypto or a libssl that used cmov (which the CPU trapped
with Illegal Instruction). Finally, we partitioned the flash drive and (with a
couple tangles not worth the space needed to explain them fully) successfully
installed and booted Debian.
We'd had a couple of ideas of how we might use the Kinect (that I happened to
still have in my backpack) to do interesting hacks, but the bar machine didn't
have USB 2.0 (only USB 1.1), so that was strictly impossible. We kinda liked
the idea of doing something with microphones, but we didn't have any. Or at
least, not what you think of as a microphone. Technically, any permanent
magnet speaker and coil of wire can act as a microphone, albeit probably with
the wrong impedance and a weak signal. marcan plugged his earbuds into the
microphone port on the motherboard, and we tinkered with settings until we
could see some difference in the data when you hollered into the earbuds.
We wrote a super-simple oscilloscope that just read in samples from the
microphone and drew them on the 80x25 character terminal in real time. It
pulsed in time with the music, or whenever anyone spoke, and really fit nicely
with the rest of the ambiance of the space station.
I'll link a video of our hack as soon as either bunnie or marcan posts it
somewhere on the net. Video below. Thanks, marcan!
Transcript:
Me: "marcan and I installed Debian on marcan's flash drive so that we could write a (um) python script to watch the audio input. But this device doesn't have any audio input so we're using marcan's headphones as a (uh) makeshift microphone. (And, uh, yeah) so we wrote this (uh)..."
marcan: "This OSCILLOSCOPE!"
me: "Awesome."
There was a girl with beautiful, incredibly long straight blonde hair - it went
down past her thighs. I asked if I could French braid it, and she let me.
Hmm. This sounds much sketchier when I write about it now than it was at the
time.
As the night dragged on, friends one by one took their leave. At 0500, marcan
and I were the last remaining from our original group, and we set out. The sun
will be up soon. (In hindsight, staying up this late is NOT going to help my
ongoing jetlag. Oops.)
I have made many friends here, and I look forward to collaborating with them in
the future. Until the next Congress, we have the internet. And in the case we
find ourselves in nearby regions, we'll meet up again.
The year in review:
Decide whether to switch degree plan to MS or not
Earn an A in Brewer's systems class
Present some piece of my work at a conference
Make friends with brilliant people whose impressive work I find inspiring (achieved multiple times!)
