|
|
|
following up on a clue
Monday, April 8 2002
"Here's a CLUE for anyone born before 1980... Blimpie is the Big One... Anything else is a BIG MISTAKE?"
Jesus I hate that commercial, particularly the way "BIG MISTAKE" is pronounced as though it comes at the end of a question. It's just one aspect of what's wrong with pop culture these days. In my opinion, it hasn't been this bad since the late-mid 1980s.
I should mention while I'm at it that the anorexic Britney Spears wanna-be in that Blimpie commercial looks to be at least 25 years of age. In other words, she herself might benefit from a few of the juicy clues of wisdom dripping down from the N'Sync-loving youth of today.
I myself followed up on a clue provided in an interesting Salon article about the study of software evolution. I got the impression from the article that computer experts are still a long way from formulating a scientific theory describing the life cycles of software, but I was also impressed by what is known. Particularly striking is the fact that, as a complex piece of code ages, incremental improvements made by developers gradually level off until an equilibrium is reached in which any additional changes introduce as many new problems as can be fixed. I don't know how true this is in the modularized world of object-oriented programming, but it is demonstrably true when dealing with spaghetti-based procedural code, the sort encouraged in such development environments as MS SQL, Assembly Language, and Commodore Basic (to name a few).
Though object-oriented, the Flash Actionscript programming environment also seems to suffer from complexity issues after about 600 lines of code. I've noticed, for example, that all the gains in code execution efficiency I'd made a few weeks ago have been reversed somehow by the process of adding more features since then. And try as I might, I cannot pinpoint the sources of any of these new inefficiencies. So today I attacked the core server polling algorithms, the basic machinery of the chat client. In any evolving software system, it's never difficult to find legacies of earlier ways for doing things, and the core polling routines were no different. But even after clearing out the cobwebs of old polling methods, the chat didn't run any better than it had. Happily, it didn't run any worse either.
For linking purposes this article's URL is:
http://asecular.com/blog.php?020408 feedback
previous | next |
