My friend Alex thought up something really great, and then, most impressively, carried it out. It was this: Mad Science Week. He's in neurobiology grad school in Boston and so of course knows myriad clever people who are usually very busy hacking on their own particular research. The idea was this: everybody take a week off of the usual grad school research to put their skillz together to accomplish a Mad Science Task, a week together engineering something fun. They did this!

Here's what Alex has to say about it:
Last week was Mad Science Week, an event that's been in the planning all summer. The basic idea is to design and execute a Tobin-esque project in a week, using collective grad school power to do something interesting. Our project consisted of a musical instrument controlled by the power of light. What this translated to was a bicycle light, a webcam, and a software synthesizer. I'll try and get a video of it out at some point, but what's important isn't the project itself, but that we finally did something. A lot of my department is lots of talk and almost no action, so it's refreshing to move past that, and spend a week programming and hooking things up and being frustrated and figuring things out together. Working hard on a project I like is infinitely more rewarding than leisure. I'm crushed that I don't get to work on a project I like all the time.
I'm flattered that he called it a Tobin-esque project, especially since I only rarely work on such things myself anymore!
Welcome to the Digital Forest!

Holy crap! The freezer trick worked! I left the dead hard disk in the freezer overnight. ... Tonight I rushed it out in a chilled pyrex dish in the company of various frozen vegetables. With power supplied, the disk just clicked twice like before. I gave it a whack and it came to life, begrudgingly, and with lots of bad scraping and grinding noises. It lived for just seven minutes, but that was enough time to copy it! (The whole disk image is only 1.4 GB!)

Thanks to Linux for making this easy:
dd if=/dev/sdb1 of=./bbsCopy entire disk partition to a file
mount -o loop ./bbs /media/loopMount that file as a filesystem
Above is a rendition (via ansi2png) of the one file I wanted from that disk: the welcome screen to the BBS I ran in junior high and high school, and, after a three-year intermission, for a brief time in college. It was called the Digital Forest (or even "Digital Forest Information System"!). The ANSI graphic was created by Lord Jazz (I never knew him by any other name) who shortly thereafter became a member of ACiD (which is apparently still around).

Maybe for hilarity purposes I'll try to boot this disk image. It's OS/2!

It's possible the freezer had nothing to do with it; when I powered the disk up earlier and got only the clicks, I didn't try hitting it or anything else.
The old splorg dot org server has been down for some time now. The domain itself has been picked up by squatters, and as for the machine--there was a snap, crackle and pop and the smell of dead electronics in the air. However, I pulled out the disks and they've sat on my shelf for a few months. Today I received in the mail an awesome little USB-to-IDE gizmo (ebay, ten bucks!) and, presto! access to my old disks again.

Some stuff recovered:(Whenever I recover long-lost files (or old journal entries!) I'm discouraged by how boring they are.)

I have another disk, this one left over from the brief reincarnation of my BBS (!), which (the reincarnation, that is) I ran at Berkeley for a month or two, just long enough for everyone to get in one last game of TradeWars 2002. That disk, however, emits a pitiful click-click when I power it up.

The Internet says I should put it in the freezer.

home again

Mar. 25th, 2006 03:09 pm

Field at Chicago O'Hare. March 24, 2006. (More pictures or journal entries.)

After two weeks away, the idea that I lived in Rochester had started to feel like the remnant of a weird dream, the kind that lingers uneasily as you ponder whether or not it really happened. The flight from Seattle was quick. The flight from Chicago a bit odd. The owner of a local Rochester basketball team was on the plane and offered everybody tickets, heightening the sensation that amongst us Rochester-bound passengers—that we're all in it together.

Thursday night we stayed again at casa de [livejournal.com profile] squarkz and in the evening went out with Danyel, adding four more bars to our Seattle repertoire. Meeting Danyel in Seattle was a manifestation of Livejournal magic. He was my TA for Paul Hilfinger's CS61B back in... was it 1999? Somehow via our various blogs we've kept in very loose contact over the years, and when he saw my post about being in Seattle he shot me an email. Now's he's happily ensconsed as a researcher for the Seattle area's favorite Evil Megacorp. At the Chapel he slapped down a laptop computer on the bar and showed us his latest work, visualizing geographically the distribution of searches on local.live.com (MSFT's take on Google Earth/Maps). Bill Gates' House and Disneyland are hot spots. I mentioned an idea for looking at Livejournal data; he said, "Hmm, maybe I will try that on the MySpace data tomorrow."

Back, as I said, in Rochester now. Time to get caught up on some things.

cracked!

Feb. 28th, 2006 05:41 pm
My desktop machine got hacked! Someone came in from Portugal and launched a scan of some MCI network... leading their network folk to come knocking real quick. I hardly use the box and it's running some old Redhat installed by my predecessor, so this should come as no surprise I guess.
I'd like to build a radio. Perhaps several. AM then FM. Receiver. It's just something I'd like to learn more about, but also I feel ridiculous with a degree putatively in Electrical Engineering but without having ever constructed any radio device more complicated than a crystal set. It seems like there ought to be a nice book out there somewhere with progressively more sophisticated designs to play with?

One little idea I had was to try to build a clock would listen to and decode the time signals broadcast by NIST; those are on nice integer fixed frequencies, so maybe something simple (with a crystal oscillator) would work? [livejournal.com profile] ioerror, Google came up with a link to some neat pictures on your flickr account in response to the query, "homemade fm radio receiver". What's the story behind this nifty little contraption?
After several days of fighting with the peculiarities of the computer system at LHO (LIGO Hanford Observatory), my analysis program is good to go—and it is awesome indeed. All I need to do is type "make submit" and my program finds the data to be analysed, constructs a tree containing thousands of processing jobs that will perform the analysis, and then submits these jobs to the computing cluster at LHO. Working late at night on a Sunday when nobody else is using the cluster, I have all of the computing resources to myself: 276 processors, each running at several gigahertz, each one with a gigabyte of RAM, with a collective total of twenty seven terabytes of hard disk space, and access to a robotic magnetic tape system containing who knows how many terabytes more. Amazingly all of this sits in one room in a remote, desert corner of Washington State. Amusingly the data processing program consists of only 216 lines of Matlab, 100 lines of Python, 26 lines of Makefile, with a pinch of sed and dash of other unix goodness thrown in for good measure. Programs that write programs: that's what makes CS fun. As I said: type "make submit", wait, then results.mat will be sitting there, ready for ingestion by Matlab and consumption by the analysis phase.

Kent and I joked that a proper scientific analysis program should start with the raw data and ultimately typeset and mail the resulting publication. I wasn't really kidding. Read paragraph three of Background: Kepler. This time I might have it: type "make analysis" and that results.mat file will be picked up and churned through the analysis. Hypothesis testing is done. Figures are generated, in PNG and EPS. The LaTeX typesetter is run, producing both postscript and PDF. If I were properly awesome, the proper concluding text will be subbed in given the results of the hypothesis testing. Ultimately: Type "make all." Robots in Washington State awaken, transcribe signals from magnetic tape to disk. An army of machines grinds through the data. A program in New York State performs model fitting, hypothesis testing. LaTeX turns this into a beautiful manuscript, which the Laserjet 2300dn fuses to the page.

Pull the page out of the printer and read what you have discovered.
I'm working on this program that be like http://www.transitinfo.org/tripplanner but for Rochester. To do this I need a list of the street addresses of all of the bus stops in Rochester and the schedules as well. I was all set to fire off a second email to the bus agency (I sent one last week but I received no response) asking for this information when I noticed their newly updated homepage:
In response to the terrorist attack on London’s public transportation system, the US Department of Homeland Security assessment raised the terror secur...
sigh... who would have thought that asking about the location of bus stops would be something likely to arouse the suspicion of the Powers that Be? I will try calling anyway and see what they say.

The urban legend that a would-be thief can pencil down your Vehicle Identification Number (VIN) and directly have a key made (and hence drive away with your prized ride) seems to be false—at least in the case of a 1992 Volkswagen. The actual protocol involves disassembling the door to get at the lock, then reading a four digit key-code off of the tumbler (whose relationship with the VIN, if any, is generally unknown), from which you can then directly cut a key. I thank AAA for doing this for me. Suckers with more "modern" conveyances, on the other hand, must have their vehicles towed to the dealer, whom they must then pay upwards of a thousand dollars for a retrofit. On the other hand, any Mark II Volkswagen can be opened instantaneously by jamming a small flathead screwdriver through the body directly underneath the lock and then thrusting upward (which I learned the hard way); with similar ease, if I'm reading the schematics correctly, the vehicle may be hotwired by bridging terminals 30 and 50 in the fusebox. Counterintuitively, the finest loss-prevention mechanism I've come across has been to leave the windows open and the locks unlocked—replacing window glass costs more than anything that could be pried out of the car.

When I was very little and learning Pascal, I had only the foggiest notion of machine code and no idea about linkers and loaders whatsoever. I could, however, write a program in Pascal and then the compiler would magically turn it into an executable (EXE!) file. The only way I could conceive to write a compiler myself was this: To write a program that was really an interpreter, and then, to "compile" some source, I would concatenate the executable of my interpreter and the source to be "compiled." The interpreter would know, when run, to examine itself, seek to the point where its own executable code ended, and find the program it was suppose to interpret. Presto! A program to generate self-running exectables from source code! But it was not, of course, in any sense a compiler.

As far as I can tell, this is pretty much how the Matlab compiler works.
For anyone who is interested in conspiring on various technical projects, such as: creating a public transporation routing program for ROC, geodesic domes, building a 'SpokePOV', micropower radio, and other small collaborative programming/electronics/engineering projects, please consider joining the Ant Hill 'Technical Projects Group' mailing list: http://splorg.org:8080/mailman/listinfo/projects (archive) There is also an associated CVS archive for real time collaboration on real code.
your livejournal username: Add this to your journal.

So, Google currently has a little promotion that involves a sort of puzzle hunt. The first puzzle, published on billboards and at [livejournal.com profile] googleblogatom is to find the first 10-digit prime number to occur in consecutive digits of e.

I tried to solve it using only standard unix programs, with the following pipeline:

cat e.2mil | delay_line 10 | factor | awk '$1=$2":" {print($2)}'

That pops out the right answer immediately...

e.2mil is a text file containing the first 2 million digits of e, and delay_line is a little C program that takes a stream of characters and whenever it receives a character after the ninth character, it outputs the previous ten characters seen. It would be much more satisfying if delay_line were implemented as a clever script and perhaps if e.2mil were generated somehow instead of with Mathematica or a Google search. (I know that there are faster ways to check primality than explicitly computing factors, but obviously this is quick enough...)

Two fun exercises on the side: download GNU coreutils and look at the sourcecode to factor to see how it works (it uses 'Wheel Factorization'), and compute the expected position of the first 10-digit prime in a random sequence of digits.

The second puzzle is to find what comes next in this sequence: 7182818284, 8182845904, 8747135266, 7427466391, _____ .

Further notes will appear on [livejournal.com profile] nibot_lab.

I spent the last two days at the Institute for Geophysics and Planetary Physics in San Diego. It was ``networking'' pure and simple, a chance for me to meet the people working there and to hear about all the cool projects that are going on. My advisor from the summer program in Alaska I did four years ago consults now for IGPP, and he invited me down for the visit.

IGPP is perched on the bluffs overlooking the ocean. You can sit outside and hack away on your laptop, connected to the wireless network, while listening to the waves crash below, looking out at the lights on the pier. Inside the discussion ranges from the formalisms of knowledge systems (and programs in prolog, haskell, and home-brew grammars for describing streams and ontologies) all the way through wireless links to sensors in the field: seismometers, digital video cameras... it's "the information superhighway meets the dirt path" or something to that effect. Cool stuff. I even ran into another Plan 9 enthusiast (who pointed me to the super cool 9grid project) — that's like meeting an Esperantist!

Tomorrow I'm visiting The Aerospace Corporation and a couple people at UCLA!

XCF

Feb. 13th, 2003 09:12 pm
Went to the XCF meeting tonight. I think I'm going to try to get involved with the xcf and to work on some kind of Plan9-related project. Met Misha Dynin there which was pretty cool. Then went to the Google-sponsored (recruitment) Pizza party, ran into Alex F. there. Altogether pretty amusing.

Oh, there was just some kind of bomb-scare around Sproul Plaza. Haven't heard anything more except that something exploded.

March 2020

S M T W T F S
1234567
891011121314
15 161718192021
22232425262728
293031    

Style Credit

Syndicate

RSS Atom
Page generated Aug. 12th, 2025 10:17 am
Powered by Dreamwidth Studios

Expand Cut Tags

No cut tags

Most Popular Tags