Acorn squash seeds

When cutting open a couple of acorn squashes for tonight’s dinner, I wondered if the seeds could be roasted like pumpkin seeds. It ends up that they can! (The same applies to other autumn squashes, such as spaghetti squashes or butternuts.)

One recipe (found with a quick Google search) is here. The commenters there say to use less oil than the recipe says – they aren’t kidding. As I did with some pumpkin seeds earlier this fall, I rinsed these a few times in a bowl of water, as I would to separate a pomegranate – it’s amazing how much gunk falls off of even relatively clean seeds.

Skein, Eric Goldberg

I’m at SIGGRAPH and SCA (the Symposium on Computer Animation) this week. Here is a doodle I did while listening to yesterday’s SCA keynote (Jeff Lieberman, who gave a really terrific talk):

(The Processing source code is here.)

I like to use these as desktop pictures.

Saturday’s keynote was from Eric Goldberg, a Disney animator. (Among many other things, he animated the Genie in Disney’s Aladdin.) He talked about hand-drawn animation, which is somewhat different from computer animation not just in technology, but in approach.

In computer animation, although shots and scenes are planned according to how they will play on the screen, construction is usually done by making a model, or series of models, which can be moved in whatever way the shots require. Goldberg, by contrast, starts drawing a character in any frame with a line of action, with the idea of what he wants to get across; everything else supports that, and importantly, can change to support that: the constraints have to do with continuity and believability, not with a model’s capabilities.

This is a nice freedom that we don’t often have in computer systems (for animation or anything else). I am not sure how computer systems might be more flexible, without losing some of the nice things about having such structured models; but at any rate there is something important in the notion of letting the idea shape the thing.

The Power of Convention

Right-to-left text—or at least right-to-left page layout, which I’ve more often run across—always takes a little extra effort for me to read and strikes me as different than what’s usual. This isn’t surprising; I grew up reading left-to-right text.

This morning I was flipping through an old National Geographic and saw a picture of machinery labeled with numbers in the following order:

(The picture was of a part of the Large Hadron Collider; the part can be seen online on the left side of the last photo on this page.)

The numbers, from one through 36, happened to mark off ten-degree increments on a circle. This struck me as odd, even though it is left to right—because I am used to reading degrees starting on the right, and following around to the left.

Convention and habit are powerful things.

Better Than Life

Yesterday, I saw this post about a new McDonald’s ad which uses a mixture of CGI and photography to show delicious-looking food. Why would you go to the trouble of modeling a burger in a computer when you could create luscious-looking (if fake) food in real life, and not have to worry about details like lighting or the condensation on the glass?

One familiar reason is of course that computer models can be easier to change, once they are built: the lettuce never wilts, the ice never melts (unless it is supposed to), and you can tweak the shot all day long. I think there is another reason as well.

Later in the day, I saw the new Harry Potter movie along with a bunch of trailers; of course those are full of terrific illusion, carefully built from parts to tell a particular story. The advertising models are like this too, as are models used in games or anywhere else.

When we build models, we emphasize certain aspects of a thing, the ones that tell a particular story. Interestingly, this is largely orthogonal to the question of realism.

(The post title is a reference to a Red Dwarf episode.)

Perlin sound

When I am writing a program that generates procedural sound, I usually start out by feeding random numbers (“white noise“) to the audio output to make sure everything is hooked up correctly. Yesterday I thought I would see what Perlin noise sounded like instead.

You can hear the results here. (Warning: please make sure your speakers are set at a reasonable level before clicking on the link.) The smoother Perlin noise makes a sound like the wind or like distant machinery.

Another author has taken a different approach to turning this kind of noise into sound, making tones when particles hit different parts of the noise. This approach brings out the chaotic, random-but-rhythmic nature of the underlying numbers.

Hello world

Richard P. Gabriel says,

Practice writing. Write every day.

I thought I’d try.