Subject: How are statistics used in computer science? On 7 May 2002 16:16:37 -0700, [EMAIL PROTECTED] (John Smith) wrote:
> What are some direct applications of statistics in this field? Queuing theory is potent. It's really hard to convince the ignoramus-programmer how bad his program actually works in a high-demand environment, if he hasn't seen some of the theory. Now that computers "cycle" a 100 million or a billion times faster than they used to, I wonder if good, new programmers automatically pay attention to how much time it takes to do stuff. It used to be essential. For quality. If you are a good programmer, I think you still do timings, somewhere; an idea of models and effective modeling will still be useful, if you are going to be good. So, you need some notion of what timing is accidental or incidental, and which too-slow performance urgently needs correction. But most programmers are not very good. There are so many who 'just get the job done' that it has left me pretty sure that there must be CS professors who never grasped (statistical) modeling for efficiency, either. I don't know that you need certain, particular courses, but Quality will require a quantitative frame of mind, and 'statistics' may help to develop that. -- Rich Ulrich, [EMAIL PROTECTED] http://www.pitt.edu/~wpilib/index.html . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
