-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On
Behalf Of Jay Warner
Sent: Wednesday, May 08, 2002 9:50 PM
To: John Smith; [EMAIL PROTECTED]
Cc: Dennis Roberts
Subject: Re: How are statistics used in computer science?
Not to disagree with Dennis' very valid point(s), but....
1) I have been told in so many words, by more than one person holding the
title of professor or ass't prof. of Computer Science, that statistics
generally, and my variety in particular (Design of Experiments stuff), are
not
very valuable for their (UG) students.
--------------------------------------------------------------------------
Jay, we have talked about many things in the past, and I have mostly agreed
with you. Here however, I agree with the sources from Computer Science, era
2002.
In the past, compiler performance was important, basic algorithms for
optimum performance, queues, input/output traffic, probabilities of
overload, reliability of software programs, etc. was of interest, and these
depended in some way or other with statistics.
Everything in the commercial world has shifted, to building layers of
software on top of foundations built in prior years. Existing compilers are
used, new compilers surface with new software such as S. Even these could
just be translators to C++ and continue to rely on the basic C++ compilers.
Microsoft Visual Basic does this, it still uses basic C++ compilers.
But this no longer is the main function of the computer science graduate. It
is building these top layers, and how they interact with corporate networks
and internet interchanges. Microsoft's .NET is just a top layer, and this is
of interest.
----------------------------------------------------------------------------
---
3) Said people were not aware of the uses of statistical methods (for
analysis of operational information) in the development of UNIX at Bell
Labs.
----------------------------------------------------------------------------
--
Unix has been developed, and now is only one of the bottom layers.
Computer Science is training students to understand current networks,
understand operating systems, understand how current under layers work, what
are all the intricacies of Microsoft Windows, all of Microsoft languages,
SQL, Windows, other languages such as HTML, JAVA, etc. and to develop
software applications for commercial use. This includes games, fast visual
displays, entertainment, interactive video, etc. The internet and its
commercial traffic, sales, databases, data mining, software to select
potential customers, etc. has been the dominate viewpoint. Not every CS
graduate works for Oracle.
The speed of processors has so increased that not a lot of attention (except
in academic quarters) is put into new compilers. We are limited by the Intel
set of instructions and CPU architecture. The outstanding exception is
parallel computer architecture and devising super fast extremely large
computers to model world wide weather (for example).
Some of my older CS texts from the 1980's had sections on statistics,
covering some of the things you mentioned. But this is 2002.
DAHeiser
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
. http://jse.stat.ncsu.edu/ .
=================================================================
