Nov 162010
 

If you tried accessing this site yesterday or today, you probably noticed that it was defaced by hackers.

The site is back now, with everything still intact.

Thank goodness for backups and the power of grep, as well as my hosting service, A Small Orange, who was very helpful in getting this site back on track.

Nov 112010
 

The Decemberists - Castaways and Cutouts - Cover Art
Album: Castaways and Cutouts

I’m heading down to Monterey this weekend for the Big Sur Half Marathon which will be a nice break after a crushing three weeks of constant work and sleeplessness. I’ve been heading down to Monterey and Big sur pretty regularly, and it’s become my favored spot for some contemplative repose.

And with this song, I’m starting to run low on songs in my collection about the Highway 1 coastline. Anyone know of any others?

The Decemberists - California One/Youth and Beauty Brigade


Buy: CD/mp3

Nov 042010
 

The Department of Justice has just filed an amicus brief supporting the upholding of the US Federal District Court’s opinion that genes by themselves are not patentable.

This is huge.

The patenting of sequences of DNA has always been controversial. Opponents fall mainly in two camps. The first argues that there are those that say DNA, as a defining element of life, should be held by a private company as a patent. I don’t agree with this viewpoint. There is nothing extrinsically different about DNA from other materials that should differentiate it by this argument.

The second camp makes a much stronger argument, and is the reason cited by the DOJ’s amicus brief, and is also the reason I am strongly opposed to DNA sequence patents. Because DNA in itself is a mere product of nature, and not in itself an invention or process or design, it is not patentable. The DOJ’s brief goes further to say that diagnostic tests using the discovered sequences are not patentable, since this is nothing more than scientific conclusion. It’s like patenting the fact that uranium decays into radium. That fact in itself should not be patentable, but if some party describes a process to detect the radium as part of a system to detect uranium decay, then that process would be patentable, but wouldn’t preclude other parties from developing potentially competing technologies based on the same scientific fact.

This NPR article cites a GW Law professor’s contention that we should extend these patents, and companies holding these patents will license out the technologies, much like computer component manufacturers license out theirs. However, this is flawed reasoning, since (1) each proprietary computer component, interface, or protocol is a very obvious invention or process, and, (2) for the key parts of the computer, the patents are placed on architectures leveraging the core materials and processes, which are shared by everyone.

My belief is that patents are critical in biotech, an industry where R&D investment for a single product can easily balloon to hundreds of million, or even billions, of dollars, and take over a decade. However, you just shouldn’t be able to patent an identified gene that appears in millions or billions of people worldwide. It’s not an invention. It’s not a process. It’s not patentable.

Oct 292010
 

The last four days has been a frentic coding frenzy–coding for 12 or more hours a day on an assignment for Stanford’s CS448B: Data Visualization course.

The final product is the Iraq War Incidents Browser. Screenshot is below (click to enlarge). Application and (pretty messy) source code, and a much more detailed description on my CS448B page.
Incident Browser Screenshot

Taking the entire data set of the recent Wikileaks leak of Iraq War Incident logs, the application can filter by date and region, and display histograms satisfying each filter.

I am willing to call this version 0.5, since the bulk of the interesting data is in the text itself.

For the eventual version 1.0, I will be adding the ability to view individual records and their content, and hopefully providing text filtering in addition time and geospatial filtering.

Go try it out!

If you have comments, feedback, or suggestions, I’d love to hear them and incorporate them into the design for version 1.0!

Update: Just to clarify, I don’t particularly believe that the release of classified logs pertaining to the war during wartime is a good move. However, since the data’s out there, we might as well try to make some sense of it and approach it in a reasonable and balanced way.

Oct 222010
 

I took a few shots of Stanford buildings last night in the dark, since the damp air along with the full moon was giving everything a nice, radiant glow. I started wondering what the fastest lenses in existence today were, and if I would ever be able to get my hands on one of them. I decided to do some searches, and here’s what I found.

Note that there are actually quite a few lenses that shoot under f/1.0, but it seems that most of them really can’t be considered sharp. The ones I looked at have the large aperture, and retain some semblance of sharpness.

I shoot Canon, so let’s examine Canon’s line first. Canon’s fastest lenses in production today are its 50mm and 85mm f/1.2. Pretty good, but of course, there’s the legendary 50mm f/1.0 that was discontinued, but is widely sought after, especially by DSLR filmmakers. Also worthy of note is the 200mm f/1.8, which, though a full stop slower than a f/1.0, is probably the more impressive given the length of the lens.

Leica Noctilux f/0.95 lensMoving away from Canon, we have the Leica Noctilux f/0.95, which sells for close to $10,000. Of course, for those of us that can afford to drop close to 5K for a Leica body and some basic lenses in the first place, this shouldn’t be too much of a stretch. Just search for “noctilux” on flicker, and you’ll find loads of gorgeous images with a DOFs that seem out-of-place for a 50mm field of view. This is probably the go-to mass-produced super-fast lens right now.

There’s also the company Noktor that released a micro-4/3 lens earlier this year in March 2010, and everyone has been following them to see if they will release a Canon or Nikon mount version of their first lens, which is a 50mm f/0.95, which retails direct for only $750. Yes, that’s about half the price of the Canon 50mm f/1.2L. But of course, the Noktor f/0.95 is nowhere near as sharp as the Canon f/1.2L, and also has some color and vignetting issues. Phillip Bloom, one of my favorite film photographers, has an interesting review of it. His posts are almost always extremely excited, so read his review with that lens. These issues aside, if speed is all you care for, then we can hope that when the EF-mount version comes out, it will be by far the cheapest of the ultra-fast lenses out there.

Zeiss 50mm f/0.7The lens that takes the cake, though, would have to be the Zeiss 50mm f/0.7. This Italian article details this lens’ history, and, if you don’t understand Italian and don’t feel like reading through badly translated Italian, this page summarizes it well. Essentially, the lens was designed as a 100mm f/1.0, with a 0.7x teleconverter that adds another 0.3 f-stops of light. With this lens, Stanley Kubrick filmed several scenes in only candlelight in the film Barry Lyndon. That (pretty incredible) task is detailed here.

So, it seems like, unless I make friends with Stanley Kubrick, my best bet is to save up for one of the Canons, or the Leica Noctilux and a converter. This is perhaps a purchase that can wait a few years or decades.

Oct 202010
 

Cover of Belle and Sebastian CD
Album: Write About Love

This past Sunday, I was at the Treasure Island Music Festival in San Francisco, headlined by Belle and Sebastian, and also featuring Ra Ra Riot, Broken Social Scene, The National, Superchunk, Rogue Wave, and other great acts. Ra Ra Riot was spectacular as usual, and Alexandra Lawn even sang a song. Matt Berninger of The National was just sipping wine the entire show and looked like the coolest frontman of all time.

This was my first time seeing Belle and Sebastian, and all the good things I’ve heard about their shows are well-deserved.

Listen:
Belle-Sebastian-I-Didnt-See-It-Coming.mp3


Get the album! CD/mp3

Oct 122010
 

One of the most covered songs ever, Feeling Good “is a song written by Anthony Newley and Leslie Bricusse for the 1964 musical The Roar of the Greasepaint—the Smell of the Crowd” (thanks Wikipedia).

The two versions here are from Muse’s Origin of Symmetry, and My Brightest Diamond from the compilation, Dark was the Night. Dark was the Night, especially is probably the best charity compilation I have ever heard, with two sides filled with great artists recording new material, and actually trying hard to create art, instead of the usual afterthoughts that appear on these type of albums.

Other notable versions include Nina Simone’s and John Coltrane’s.

Listen:
Feeling Good - Muse
Feeling Good - My Brightest Diamond


Get Origin of Symmetry! CD/mp3
Get Dark Was the Night! CD/mp3