job switch / google

I now work for Google. Hooked was great fun, and I worked on interesting projects with wonderful people.

Ames never ceases to astound

I love how "far out" / "long range" some of NASA Ames' research is, for instance :
http://www.kurzweilai.net/nasa-ames-worden-reveals-darpa-funded-hundred-year-starship-program in which the authors discuss a new joint DARPA/NASA Ames Research Center project to build an "100 year starship"

View of the San Joaquin Valley, including fruit/nut trees and aqueduct

blogger question

Does anyone know if there is a way to get the "summary" field in the atom feed from a blogger blog to include thumbnails of the images in the blog?

I recognize that this would probably require some coding on Google's part, but it doesn't seem like anything that's beyond them, and it would be a pretty cool feature/option.

Hyperbolic Paraboloid Buildings

Ever since I first stumbled upon them in "A Visual Dictionary of Architecture" I've been fascinated with the concept of hyperbolic paraboloid buildings.

Essentially a hyperbolic paraboloid is a structure which, in some frame, is described by the equation

z=x*y

If you slice it along constant z you get

1~ x*y or y~1/x or x~1/y

If you slice it along a line y=k*x you get

z=k*x*x or z=k*x²

But the part that makes it useful as a surface for buildings is that if you slice it along constant x or constant y you get

z~x or z~y

in other words, a straight line.

For this reason, it is called a "ruled surface", and a framework for it can be made entirely out of straight beams, as in the pictures from this site

http://www.savetrees.org/Hyperbolic%20Paraboloid%20roof%20shelter.htm



One problem with such a surface is that, while it is easy to construct the frame out of straight beams (and the resulting structure is known for its strength), constructing the roof to fit into the frame is non-trivial, especially if the surface must be hard.

It is *possible* to construct something that mostly fits snugly over it by cutting pieces out of a cloth tarp (as shown in the above link), but any patch of the true surface of the object is curved (as well as any line which is not constant in either x or y)

One solution might be a "concrete tent" similar to what is proposed in this article
http://www.wired.com/science/discoveries/news/2005/03/66872 (start with a "tarp with sections cut out" and brush wet concrete over it. Then let the structure harden)

Another alternative might be to layer a fine mesh of straight wires along constant x and constant y, and take some goopy filler material, such as plaster, and brush it over the wire mesh. I wonder if something like this is how they make permanent buildings with hyperbolic paraboloid roofs. Some of the images of such structures certainly look like a wet material was painted on top of a mesh and then allowed to dry, although larger hyperbolic paraboloid structures, such as the Catholic Cathedral on Gough Street in San Francisco, are often clearly made from piecewise surfaces (float or otherwise). I think to have a piecewise surface of flat segments perfectly conform to a frame of straight segments in such a roof, the segments have to be triangles (and highly irregular ones at that!)

This time last year

This time last year, the mountains around Los Angeles were on fire. At 15 miles away, you could see the flames sometimes in broad daylight. And at night a 30 degree arc of the skyline would glow red. One of my favorite nights in Los Angeles was the time I sat out with a bottle of cheap red wine just watching as the sky burned around me.

Side note : the way I found the time-lapse video of the high voltage tower construction
http://mikeschuresko.blogspot.com/2010/01/really-cool-time-lapse-video.html that I posted around this time last year was by looking for videos of the Station Fire, and finding this one
http://www.youtube.com/watch?feature=iv&v=jR_3N7nVPw8 by the same guy. I like the music in his videos, but I feel that they don't adequately capture the scale of the fires or how the sky turned red from the other side of the city or how everything smelled like a smoky oak campfire.

Planning

On my bike ride to work, from Mountain View to Cupertino, I usually start on the Stevens Creek Trail, and I can take either one of two routes after crossing El Camino Real.

One of them takes me through the red circled region on the map posted below, while the other takes me through the blue.



I will call the blue region (the "circle" around it isn't closed since the region extends off the map) "ecocity Sunnyvale". It has actual bike lanes, streets connect, so that things a close distance apart can actually be walked or biked to, and the area seems to be fairly mixed-use (schools, businesses, etc) for a residential / sprawling ranch houses area.

The red region, while superficially similar by the map, is an entirely different beast. The dashed green line going through it on the bicycle map isn't so much a bike path, as it is an unmarked route through a maze of suburban cul-de-sacs. Having successfully laid out the streets in such a way as to prevent cars from cutting through this neighborhood, whoever designed this area also managed to prevent cyclists and pedestrians from taking any sort of sensible straight cut-across, while simultaneously zoning the entire region purely residential.

Re-posting an interesting idea.

Re-posting
http://robotmonkeys.net/2010/07/22/bullet-train-hopping/

Jonathan writes:

"Jianjun Chen in China proposed an interesting idea for eliminating station dwell times for trains. In his/her design, each train has a detachable boarding shuttle mounted on the roof of the train. Passengers who wish to disembark leave the main passenger compartment of the train, and enter the shuttle. Meanwhile, embarking passengers board an identical shuttle already located at the station. As the train approaches, the shuttle mounted on the train, disengages so it can slow to a stop at the station, while the shuttle is grabbed and mounted onto the moving train."

http://earthobservatory.nasa.gov/IOTD/view.php?id=44717&src=eorss-iotd

You can see Santa Cruz (and the rest of redwood country) on this map.

A manifesto in the form of a sequence of seemingly unrelated ideas

• Programming is a social activity. Sure there are cases where you are working with a fixed set of libraries you know well, and you're trying to generate something interesting within this vocabulary of thought, but most industrial coding seems to consist of cobbling various libraries together.
  
• Computing is a social activity. (no explanation needed)
• (completely unrelated) Things that only the "dorky kids" did in my generation are the exact things "all the kids" seem to be doing in subsequent generations.
• Programming is the fundamental activity people do with computers, much like "driving" is the fundamental activity people do with cars
  
• Learning to program only *seems* hard because we try to force students from zero to programming-literacy in one college semester, then fail them if they fall behind on the interesting stuff that depends on that basic literacy. Imagine if we spent a year slowly teaching programming for every year we taught kids about reading, writing or arithmetic.
• Analogy between modern professional programmers and ancient scribes.
  
  • Elite group of educated scholars
  • Write using needlessly difficult technology (either "C" or "All caps and no punctuation")
  • Mostly write things like "So and so owes the king 15 sheep" / modern day business software
• Analogy between programming and other basic intellectual tasks.
  
  • Perhaps in the future being a "professional programmer" will be as weird as being a "professional writer" or a "professional mathematician" today
  • But, likewise, everyone will need to do a little reading/programming/arithmetic
  • There may be many "high school programming teachers"
• Printing press analogies
  
  • Looking at computers and the internet to day and getting interested in them as "fascinating machines" is like looking at the first printing presses and becoming obsessed with the screw mechanism.
    
  • Omitting the rest of the analogy

San Francisco bay from Mountain View

Robot household cleanup and RFID

What if every thing in your house that wasn't trash had an RFID tag?

Your household swarm of cleanup robots could potentially pick up every item. If the item doesn't come with an RFID tag, it goes in the trash/recycling/etc. (telling these apart might be hard). If it does come with an RFID tag, there's a household database that tells your robots where the thing goes when it gets put away.

I suppose this wouldn't solve all the problems with "program robots to clean your house" but it sure seems like a big first step. (It might be worth testing any proposed solution to the "robots clean your house" problem with the "the baby just defecated on the floor" thought experiment)

(food items would be an interesting issue, the RFID tag would have to have some sort of expiration information, and when the food gets thrown out / composted/recycled, the RFID tag has to be removed for later reuse. Things like shampoo bottles and toothpaste tubes present similar issues.)

P.S. If any readers find similar ideas posted elsewhere, please post links to the more interesting write-ups in the comments section.

The right approach, I think

http://www.oxygen.lcs.mit.edu/Overview.html

Having glanced through the websites of several "ubiquitous computing" research groups, I think that the Oxygen project at MIT is the closest I've seen to "something on the right track" Partly I occasionally track / glance at what's happening in this field because I think it could become an interesting thing to be involved in in the future. Partly I keep looking at it because I have this sneaking suspicion that, when we start getting to pervasively embedded small computers, HCI will start looking like a robotics problem. And partly I'm interested in how the network connectivity for pervasive computing might work : perhaps the distributed robotics and ad-hoc network people have a head-start on some of the issues that might come up.

Should be of interest to some

Particularly of interest is the discussion of the "software as object" focus

http://www.shirky.com/writings/group_user.html

Group as User: Flaming and the Design of Social Software

Central Valley Airport

Why I think it would be reasonable to construct an international airport in California's Central Valley (pending the construction of at least the Central Valley portion of the California High-Speed Rail project)

1. Its flat (so you can arbitrarily scale it up with new runways) Of course the current solution to overcrowding at SFO and LAX is to increase runway space at outlying airports like SJC and Ontario -- one of which is already in the central valley, while the other one is along one of the proposed high-speed rail lines (albeit the part of the line least likely to be built soon -- most sensible discussion thinks high speed rail will connect to CalTrain/Metrolink, and CalTrain will take care of the San Jose --> San Francisco leg)
2. High speed rail will make a hypothetical Central Valley Airport accessible from the current major metropolitan areas (not useful if you're flying to LA, but what's an extra hour of train ride if you're going to Singapore?)
3. One of the parts of the Schwarzenegger agenda I actually sort-of-agree with is the (unstated) intent to push urban/suburban growth into the Central Valley, most likely along the CA-99 corridor. The housing bubble stopped this somewhat, but the infrastructure/empty housing is in place, and a chain of UC campuses stretching from Davis to Riverside could spur job growth.
4. Part of me secretly wants to teach at UC Merced. I predict that, as time goes on, UC Merced will rapidly become a top engineering school, (making it even harder for me to do this). High speed rail and a massive airport would make it easier for an urbanite like me to live there.
5. Unrelated, but I'd like to curse the Southern California software industry for locating primarily in the public transit deadzone of Santa Monica. Although I'm told that there is a new Expo Line they are building between downtown and Santa Monica. I'd like to remind everyone that the massive Southern California megalopolis is "train scale" and not "car scale."

(edited later to add)
One of the drawbacks to this plan (and plans involving "Central Valley growth" in general) is that the Central Valley is already being used for agriculture (and is fairly unique agricultural land). According to Wikipedia,

"The Central Valley is one of the world's most productive agricultural regions. On less than 1 percent of the total farmland in the United States, the Central Valley produces 8 percent of the nation’s agricultural output by value: 17 billion USD in 2002."

Presumably some of it can be diverted from agricultural use, but it could radically change our food production if it were to entirely convert over to sprawl.

On band.

I was sitting home sick, watching youtube videos

http://www.youtube.com/watch?v=qybUFnY7Y8w
http://www.youtube.com/watch?v=PxANVR07GVI

(who does their videos, by the way? Pete would probably know.)

when I came across
http://www.youtube.com/watch?v=EltW4w2W_Ys and I started thinking about my "high school band" days, and how our sound always sounded so *mushy* and I had the recurring thought "If the band's seating is 100 ft across, and sound travels at 1000 ft/sec, the clarinets should hear the trombones be off by 1/10th of a second -- or, at 120 bpm, 1/5 of a beat."

I mean, I'm sure skilled musicians (not me at the time) compensate for this in a variety of ways : but any form of compensation you could have only "works perfectly" for at most one spot (2 spots if the band is aligned precisely along a line). It seems like the "best" you could do is "everyone plays off the conductor's baton" (information from this travels at the speed of light) and for a listener at a large distance, l, from the center of the band, to whom the band is aligned perpendicularly, hears different instruments as if they were at most

sqrt(l2 + 2500) - l

feet apart, which, as l gets large, behaves sort-of like (1/(2*l)])*2500. Or, better yet, pick the spot (the king's chair?) at which the sound has to be perfect, and align the entire band along the rim of a semicircle with that spot as the center (somewhat close to what bands and orchestras actually do).

But still, the time delay to to sound travel when you're sitting inside the large band/orchestra means that musicians can't really react to each others sound they would in a small ensemble. In essence, they have to rely on the (centralized?) conductor for synchronization. Which is probably yet another one of the many reasons why groups that do real improvisation tend to be small.

Also, that guy with the monocle in the elite box off to the side? He's probably hearing one of the worst orchestral performances of his life.

I have a weakness for far-off space station designs

Cool

http://lifeboat.com/arki/design-construction.shtml

I'd live in one. I like the whole modular-architecture aspect to it as well.

Not as expansive or ambitious, however, as :
http://en.wikipedia.org/wiki/Island_Three

OpenSocial etc.

A thought about where "social networking" could go : instead of just being "a thing that lets you post messages for your friends", the permission groups on social networking could be "the equivalent of AFS-style permissions and groups on the internet-wide massively parallel computer"

If this is combined with a shift of UI metaphors from "desktop, file, etc" to activity and shared artifact it could provide the mechanism of access control for all of the things we do on computers once we start wanting to share not only the finished products, but also edits, revisions, etc, with friends.

really cool time-lapse video

http://www.youtube.com/watch?v=iTfx9n2ImMc


Timelapse video of construction of a high-voltage tower in Southern California / Central Valley

I think I should probably learn this...

https://cvs.khronos.org/svn/repos/registry/trunk/public/webgl/doc/spec/WebGL-spec.html

WebGL. Apparently it integrates with the "canvas" element. I wonder if IE will properly support it.

I wonder if it matters anymore at this point whether IE supports it ;)

Seems to me, though, that the bulk of the applications I use either run or can run over the web, that its easier to get someone else to run your web app than your "compiled for XXX platform" app, and that Javascript/HTML/CSS is almost entirely unsuited for the type of applications I'd want to write. So, on that side, its probably about time for something like this.

On the other hand, I seem to recall VRML having Javascript support way back in the day, and am somewhat inclined to describe "manipulating scenegraphs with a functional language" as a more advanced form of programming than "mimicking C-style OpenGL calls in the same functional language" It is not as if one could easily implement an efficient scenegraph on top of a JavaScript OpenGL API. On the other hand, this sort of thing *might* be the appropriate way to "incorporate 3d style effects into a 2d Javascript UI"...