|
|
|
fruitless experiments and debugging
Tuesday, March 26 2002
Because of the peculiarities of my attention span and the things that interest me, my projects (be they paintings, computer programs, or musical projects) usually follow a somewhat predictable trajectory. They start out with a grandiose slap-dash effort that attempts to establish something complete in and of itself in a minimal amount of time, preferably an hour or so. Depending on my motivation and the possibilities, at that point I either have a finished "work" or merely a starting point from which to see what might be added or improved.
The Flash chat project started out two weeks ago as a simple proof of concept. At first I had simple questions such as, "Can Flash read an XML document that doesn't have a .XML extension?" These were all satisfactorially answered in the course of only an hour of experiments, at which point I'd already built the core for the chat. From then on, the chat project was like a snowball rolling down a hill, a self-motivating project of incremental feature changes and occasional rewrites of the core functions. Since I am still fairly new to Flash, though, there was also a continual need for experimentation and debugging. It's probably safe to say that fruitless experiments and debugging occupied at least 98% of my chat development efforts over the past two weeks.
Today my principal discovery was that my client-side memory leaks were mostly the consequence of overly-rapid polling of the backend. Slowing this polling down by 66% or so resulted in an almost complete disappearance of CPU and memory usage creep. According to Bathtubgirl's server guy Sean, there had also been similar resource problems on the backend, but with the slower polling rate, they vanished as well. This development was very important, because the resource issues had been severe enough to threaten the chat's viability as a deployable technology.
Tonight Gretchen and I were talking about whether or not assassination could make a positive change in a country's political structure. I said that I didn't think it could in the United States, since no particular elected position has all that much power. This fact, coupled with the martyr effect means that an assassination generally helps the causes of those assassinated in this country. For example, it seems that most of the gains of the Civil Rights movement weren't consolidated until after the assassination of Martin Luther King, Jr. Similarly, the assassination of John F. Kennedy probably played a decisive role in the passage of the original Civil Rights Act and the continuation of America's Lunar missions. "But what about the assassination of Hitler?" Gretchen asked. I thought about it for a moment and agreed that in some cases (mostly in other countries), individuals are so powerful and so crazy that their assassination can bring about improvements.
Then, in an effort to fully explore my feelings about assassination, Gretchen turned the tables and asked me what I would do if I came upon a helpless George W. Bush drowning in a river (or, more realistically, passed out from a cocaine overdose). "I'd probably rescue him," I said, explaining that the martyr effect even benefits those who die from their own stupidity. "But what would you say to your friends when they asked you why you saved him?" she asked. "I'd tell them all about the martyr effect, and how my saving George W. Bush probably helped prevent truly horrendous pro-Big-Energy legislation from sailing through the Congress."
About the only place in America's politics where a positive change could result from assassination is in the case of the Supreme Court. It's hard to imagine even George W. Bush coming up with a worse replacement for Clarence Thomas. Nonetheless, I wouldn't want to risk it. Similarly, it would be most unwise for the Isrælis to allow any harm to befall Yasser Arafat. Compared to many Palestinians, he's a reasonable man.
For linking purposes this article's URL is:
http://asecular.com/blog.php?020326 feedback
previous | next |
