At work, a big part of my job is system optimization - making things run as quickly and as efficiently as possible (can't always have both at the same time). Sometimes it's necessary to tell the users that although something would be nice to have, it might be impractical.
The last post about time zones made me think of this. Pixy, when you do these magic fixes or patches, how much does it cost in terms of overall system speed? Are there certain kinds of magic that aren't worth the overhead?
I work on mainframes, so this PC stuff is pure curiousness to me.
Posted by Ted at April 20, 2004 03:13 AMHey Ted, I think you forgot to put in a Geek Alert as I that was way over my head and I was hopping up there at the time. 
A big part of my job has always been system optimization too, so I always test the performance of anything I put in place. The thingy I'm planning to do to fix the timezone problem exports the data from MT, runs a little Python program to sort and format it, and puts it in an include file for anyone to use. Once the data is out of MT, processing takes a few milliseconds.
That's why I've been so annoyed at the performance of blogging software (particularly MT, but other systems as well). This server can do well over 3 billion instructions per second, so if anything is slow it's because either (a) you're doing someting really really big or complex or (b) your code sucks.
As a rule of thumb, you can safely assume (b).
Posted by Pixy Misa at April 20, 2004 08:33 AM