home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.edu:2192 comp.lang.misc:4018 comp.software-eng:4910
- Newsgroups: comp.edu,comp.lang.misc,comp.software-eng
- Path: sparky!uunet!noc.near.net!mv!jlc!john
- From: john@jlc.mv.com (John Leslie)
- Subject: Re: Programming language evaluation criteria
- Message-ID: <1992Dec13.011234.4738@jlc.mv.com>
- Followup-To: comp.edu
- Organization: John Leslie Consulting, Milford NH
- References: <1992Dec12.091818.1847@msus1.msus.edu>
- Date: Sun, 13 Dec 1992 01:12:34 GMT
- Lines: 48
-
- jhowatt@eeyore.stcloud.msus.edu (Jim Howatt) writes:
-
- > I'm teaching a comparative programming languages class next quarter using
- > Sebesta's second edition as the primary text.
-
- I've never read Sebesta, first or second; and frankly I'm not about to
- drop things to read it. If Jim needs responses particular to it, he'd
- best post some specifics.
-
- > In the first chapter he lists several language selection/evaluation
- > criteria; but the list "feels" uncomfortably incomplete and subjective.
-
- In the real world, the list of criteria has one item -- politics.
-
- > In view of the flame wars that we often see among proponents of
- > different languages, I'd like to present my students with a set of
- > reasonably objective criteria for selecting a language for a given task.
-
- Why?
-
- > I've already been thru a half-dozen programming language texts and many
- > articles, but think that there must be more criteria to consider, especially
- > from practitioners who've had to evaluate and select languages for real
- > projects.
-
- I've seen literally hundreds of criteria, most often generated after
- the fact to justify a (usually reasonable) choice already made. I think
- such a list is nearly useless, unless it has stood the fire of flame-wars
- on the Net. I _would_ welcome some submissions, open to flaming, because
- we _might_ be ready to reach consensus on a few criteria...
-
- > I'd appreciate any insight and pointers to insight that you may have.
-
- I really think you're barking up the wrong tree. The first thing a
- course on comparative languages should teach is that most algorithms can
- be programmed in any of them. The point _shouldn't_ be that Algorithm A
- is best programmed in Language L; rather students should get a feeling
- for tradeoffs -- Language Q may not be ideal for Algorithm A, but the
- penalty is in the neighborhood of H hours, which might be made up elsewhere
- in the project.
-
- Students should leave your course with a feeling that most times the
- savings from choosing the right language are less than the time they'd
- spend fighting the political battle to use it.
-
- John Leslie <john@jlc.mv.com>
-
- PS: I've set followup to comp.edu, where I think the discussion belongs.
-