home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!newsstand.cit.cornell.edu!vax5.cit.cornell.edu!horj
- From: horj@vax5.cit.cornell.edu (Eric Scheirer)
- Newsgroups: sci.logic
- Subject: Computer Science vs. Software Eng. (was re: Natural Kinds)
- Message-ID: <1992Nov12.222503.15918@vax5.cit.cornell.edu>
- Date: 12 Nov 92 22:25:03 EST
- References: <1992Nov3.214913.25344@dcs.qmw.ac.uk> <MARTINC.92Nov9095218@grover.cs.unc.edu>
- Followup-To: sci.logic
- Distribution: sci
- Organization: Cornell University
- Lines: 44
-
- In article <MARTINC.92Nov9095218@grover.cs.unc.edu>,
- martinc@grover.cs.unc.edu (Charles R. Martin) writes:
-
- > engineering vs. chemistry, there are techniques and methods that apply
- > to software engineering that are *used* in computer science, and
- > (naturally) methods and techniques from computer science that must be
- > used by software engineers.
- >
- > As to whether they *should* be separate: of course not. All software
- > engineers know that computer science is a subdiscipline of software
- > engineering.
- > --
-
- I've been following this thread with great interest, and thought that
- it might be time to inject my bit. I'll specifically disclaim my
- qualifications here; I'm an undergrad in Cornell CS, where I've been
- focussing on theory. I could probably handwave answers to most of
- Prof. Pratt's lovely list of comp questions other than the graphics/image
- processing ones. I also have a few years' experience as a C grunt in
- a programming house.
-
- I think that an interesting comparison of science vs. engineering vs.
- math in computing science can be drawn around the lines of program correctness,
- or verification (Q. #12).
-
- There are, empirically, at least three major ways of detecting bugs:
-
- 1. Use the program until it fails to work.
-
- 2. Try to construct a lot of test cases during development.
-
- 3. Try to prove the program correct (in the Gries sense) before coding it,
- and fail.
-
- Certainly, it would be possible to construct an argument that these
- methods correspond to "science", "engineering", and "math" respectively.
-
- Program correctness seems to have a wide range of approaches, which
- seem to cover the field from the clearly "math" to the clearly "engineering".
- --
- Eric Scheirer -- Cornell University, Ithaca, NY -- (607) 277-1899
- "I got remote control and a color TV; I don't change channels so they
- must've changed me." -- Billy Joel
-
-