Linkstew

Yet more evidence that I have altogether too much time on my hands:

When a reader clicks on the "Random" link over there on the side, they're taken to a random entry in the stew. But as it is, I have no way of knowing which entry a reader randomly sees. When I notice a random hit, I often wonder about which entry a reader randomly ended up at.

But today, I saw someone click the random link, and then saw them follow several links from that random page to other pages within the stew. And, having too much time on my hands, I took destinations and triangulated back to which random entry the links had originated from.

I really need to toss an image that includes the random entry number being displayed as an http parameter on that random page. Y'know, a nice, simple, <img src="random_tracker.gif?noid=entrynumber"> would do the trick.

I'm a little embarrassed to admit this, but it took me about 30 seconds longer than it should have to figure out what was "wireless" about my co-worker's "One Subject Wireless Notebook."

Today I renamed my work notebook from "triage" to "bsii's virtual memory," because it's where I put things I can't think about immediately.

This isn't good. And considering that it was in the process of installing the AppleScript update when this happened, this isn't good at all.

The system hasn't been affected? My foot! Now I'm afraid to reboot.

I couldn't sleep, so on a whim, I started doing sit-ups.

Lesson 1: Doing sit-ups in bed is easier than on the floor, because you bounce.

But then I tried to do some push-ups.

Lesson 2: Doing push-ups in bed is much harder, because the bed absorbs a lot of the force you're trying to push yourself up with.

One of the classic questions of cognitive science is "Why do things look as they do?" Two opposing answers to this question are "Things look as they do because we are the way we are," and "Things look as they do because the world is the way it is."

On the one hand, things look as they do because the world is the way it is. Photons are emitted by light sources, bounce off of things, and into our eyes. The wavelengths of these photons determines the colors we perceive.

But on the other hand, things look as they do because we are the way we are. Any optical illusion is evidence of this, because in an optical illusion we perceive something that is contrary to what exists in the world.

The Cognitive Science classes I've taken here at Berkeley concede the necessity of environmental stimulus, but beyond that are pretty organism-centric. The phenomenon of optical illusions (among others) is taken as evidence that our experiences of the world are dependent on our biology.

But "because we are the way we are" always seemed to be missing something for me, and today I figured out what it was.

Why do things look as they do?

Because we are the way we are because the world is the way it is.

We see the colors we do because our eyes are visible to a specific wavelength of light. Why are our eyes sensitive to visible light instead of infrared? Because the majority of the light that reaches the surface of the earth is in that wavelength. Why does the wagon wheel illusion occur? Because the proximity heuristic used to help solve the motion correspondence problem was more accurate than other failed (and hence unknown) heuristics, which gave the organisms with those heuristics an advantage.

If our species had developed on a world where gravity behaved differently, or where other wavelengths of light predominated, we would be different organisms, and we would most likely experience the world differently. Our brains were shaped by the environment in which they developed, making our experiences just as much a product of our environment as they are a product of our biology.

(Please note that I am not talking about "nature versus nurture." That's a whole other can of worms that I'm not even going to get into.)

Ack! My laptop's taken a direct hit to her starboard battery! She's venting drive plasma!

That is to say (in non pirate-trek), my laptop's second battery just died, and now refuses to charge at all. For the last couple of months it would only last about 20 minutes anyway... At least I had two. September/October (when I hope to buy a TiBook) can't get here soon enough.

It's kind of weird to not have my starboard right palm resting on a warm surface anymore, though.

Inspiration struck me tonight, and realized why I blog: I'm backing myself up.

Now stick with me for a moment while I explain this, because there were a number of steps involved in my reaching this conclusion.

When I find myself explaining the difference between "RAM" and "hard disk storage" to someone who doesn't really know computers, I always end up using the metaphor Computer Memory is Human Memory to help him understand things.

"RAM is like a person's short term working memory he uses when he's actively thinking about something," I tell him. "RAM is that 7+-2 you hear so much about. RAM is where your computer keeps what you're typing until you save the changes, just like short term memory is where your brain keeps what you're thinking about until you make the effort to 'commit it to memory.'"

"Hard disk storage, on the other hand, is like where your brain keeps your memory of your first kiss," I continue. "You have access to it when you want it, and the rest of the time it just sits in your brain until you need it. When you hit 'save,' the hard disk is the memory to which what you typed is committed."

"A person can only think about a few things at once with his short term memory, but by buying more RAM, your computer can think about more things at once. And a person can remember a lot more than he can think about, and by buying a bigger hard disk, your computer can remember a lot more." I explain.

But in Microserfs, a comment is made that "We've reached a critical mass point where the amount of memory we have externalized in books and databases (to name but a few sources) now exceeds the amount of memory contained within our collective biological bodies," and that statement stopped me in my tracks.

In a flash of insight, I turned the metaphor Computer Memory is Human Memory around, switched Human Memory to Human Knowledge, and ended up with Human Knowledge is Computer Memory.

Looking at it from that direction, I realized how amazingly volatile and short term all of human memory really is. We don't "save" things to our "long term" memory. Long term memory happens to be a little less volatile than working memory, but both of them will be lost if we crash.

I realized that if I don't externalize myself, then when I "shut down," my unsaved changes will be lost. I blog in order to save me. I'm backing myself up so that once I'm turned off, my data will still live on.

Through the looking glass of Human Knowledge is Computer Memory, my short term memory becomes registers for my CPU, and maybe a little cache, while my "long term" memory is just my main memory. It's all volatile, though. My blog is my hard disk.

And sprinkling in a little Douglas Adams for good measure, the human species becomes a massively parallel processing computer. Every human experience provides our internal computers with more data that, in combination with data each one of us has acquired by accessing stored human knowledge, produces more data that can be added to the human experience.

But to be part of the human experience, that information has to be saved, by blogging (or other means). Catalogued and indexed in computers, human experience is preserved and available for future generations to build upon, eventually synthesizing the question whose answer is 42.

That last paragraph is awfully human centric, but it's humans who save their changes for later humans to synthesize into still more complex ideas. Like this one.

So one of my biggest issues with Episode II was how disjointed it felt. Every time I started to feel like a scene was being established, it was gone. It just didn't flow. And whether that was the fault of too many characters doing too many things in too many places, or the fault of bad directing, I won't speculate.

But Episode II's problems pointed me at a more troublesome question: How on earth is Peter Jackson going to handle the transitions in The Two Towers? The Fellowship of the Ring flowed very easily and obviously as a movie, because they were all together. But the fellowship has been broken, and now the screen time is going to have to be split between three different parties that have practically no communication for the duration of the story.

So I've got a few ideas about how to handle to handle The Two Towers:

• The Two Towers: A Tale in 3 Acts. Theaters will love the intermissions, because they'll triple the opportunity to rip off moviegoers at the snack bar.
• A highlights reel, Olympics style, for those folks who thought the first movie was too long.
• When Frodo isn't on the screen, the other characters should be sad, look around, and say "Where's Frodo?"

My roommate finally saw Episode II this weekend, and he offered up this follow-up quote:

"Not even muting could save this."
-- my roommate, on Episode II's love scenes.

On a related note, I'd like to point out that Lucas is at least consistently ridiculous in his scripting of Anakin:

In Episode I, Anakin asked Padme "Are you an angel. I've heard the deep space pilots talk about them. They live on the moons of Iego, I think. They're the most beautiful creatures in the universe."

Meanwhile, in Episode II, Anakin says to Padme "I don’t like sand. It’s coarse and rough and irritating and it gets everywhere. Not like here. Here everything is soft. And smooth."

I finally got around to fixing Linkstew's slowness.

So what was wrong?

I started by putting in timing code to see how long each piece of the page was taking to render, and I found that all of the hang up was in building the central list of posts. Further poking revealed that it was the sql query itself that was taking anywhere from 4 to 10 seconds to execute (depending on mysql's mood), and this was the source of the long loading delays the site had been exhibiting.

Now, I hadn't changed that query in months, so I still can't account for why Linktew suddenly got so slow last month. But using the "explain" SQL command, I was pretty quickly able to deduce the source of my problem:

My comments table has a column that keeps track of the post number that any given comment is associated with. However, the entry number column (which I was joining my posts table to my comments table on) wasn't indexed. So to display the comment count for any given entry, my query had to do a full, unindexed scan of my comments table in order to find out how many comments any given entry had. Multiply this by every entry on Linkstew, and I had one expensive operation on my hand.

A simple little alter table comments add index ... fixed the problem good, and the query that was taking 4 to 10 seconds now takes 0.15 to 0.30 seconds.

Which means that Linkstew is once again faster than Fury.com, and all is again right in the world. ;-)

(I'm just trying to make Kevin's life difficult. He knows I love him.)

From Programming Perl, 3rd Edition:

unsift
unshift ARRAY, LIST


This function does the opposite of shift. (Or the opposite of push, depending on how you look at it.) ...

And though that's true, that's a really problematic description, because in those terms, pop also does the opposite of shift. (Or the opposite of push, depending on how you look at it.) And knowing the difference between pop and unshift is kind of important.

Luckilly, the book does't describe pop in those terms, and the book's unsift description goes on to give an example of exactly what it does.