NetNews Usenet Archive 1992 #30/NN_1992_30.iso/spool/comp/object/4557

Text File | 1992-12-16 | 3.5 KB | 67 lines

Newsgroups: comp.object
Path: sparky!uunet!spool.mu.edu!darwin.sura.net!news.duc.auburn.edu!eng.auburn.edu!henley
From: [e-mail] (James Paul Henley)
Subject: Re: Is Borland the leader in technology?
Message-ID: <[e-mail]>
Keywords: engineering
Sender: [e-mail] (News Account)
Nntp-Posting-Host: einstein.che.eng.auburn.edu
Organization: Auburn University Engineering
References: <[e-mail]> <[e-mail]>
Date: Wed, 16 Dec 1992 16:19:42 GMT
Lines: 53

In article <[e-mail]> [e-mail] (Phillip Lindsay) writes:
>I don't think my sentence implied that "we didn't need high level tools?"  I
>was merely expressing the point that high level tools don't address all
>issues.  For instance, a friend of mine (with a Computer Science background)
>was working with several PHDs on an SDI project.  The scientist's speciality
>was material science. They used "high level tools" on Sun Workstations
>to design silicon and needed my friend's help to build tools which
>understood the silicon and could test it with several tools and models in
>a reasonable time frame (with the help of a Sun hosted number cruncher).
>The project could not have been completed without his expertise.  The funny
>thing is that he ended up designing silicon and becoming a department expert
>before budget cuts chopped the department.
>
>--
>Phillip Lindsay "Government for the Politicians for Japan."
>INTERNET: [e-mail] ** UUCP: uunet!zardox!dhw68k!arasta!philbo
>Phone: [phone] ** USMAIL: 2955 Champion Way, #175, Tustin, Ca. 92680


I am looking at this from both the top and the bottom.  In order for the
tools to be effective, someone has to train Chemical Engineers in the proper
use of the tools.  That is  *my* job.  We have to cram a lot into those
two programming courses.  We are able to cram a lot more in with Turbo Pascal
than we could have ever dreamed of using FORTRAN. I'm looking at what the
next step might be.  We will always teach a procedural high level language,
because one of the goals of the programming course is to teach students to
think logically about a problem solution.  In fact, the guidelines from
CACHE state that a computer program is a formal statement of the solution
of a problem.  But in the upper level courses like Process Control, it is
impossible to teach enough in a traditional procedural language in time for
them to write the necessary programs.  We have gone back and forth between
several different simulation packages, and we are starting to use ASPEN on
Sun stations.  We have real fancy graphical interfaces, but IMHO, what is
beneath the interface is inefficient and inadequate.  Yet I'm told that it
is "State of the Art."

It will take a critical mass of graduating engineers, trained in better
computer methods to budge the status quo in industry.  I want a methodology
that can be taught at the undergraduate level, and will be powerful enough
to allow companies to do what they cannot do right now.  Probably the
most important thing is that it must allow an engineer to describe a problem
in terms that he understands, and in a language that is as straightforward
as possible.

Why OOP?  We already use object oriented methods, we just don't call it that.
We call it Unit Operations.  But going from the description of the process
to a program for implementing a model of the process is a quantum leap.


Dr. James P. Henley Jr.
Department of Chemical Engineering
Auburn University