home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!haven.umd.edu!darwin.sura.net!mips!swrinde!network.ucsd.edu!lyapunov.ucsd.edu!mbk
- From: mbk@lyapunov.ucsd.edu (Matt Kennel)
- Newsgroups: comp.lang.misc
- Subject: Re: Small Language Wanted
- Message-ID: <173ksiINNiov@network.ucsd.edu>
- Date: 21 Aug 92 20:49:22 GMT
- References: <1992Aug21.140111.24487@mintaka.lcs.mit.edu>
- Organization: Institute For Nonlinear Science, UCSD
- Lines: 66
- NNTP-Posting-Host: lyapunov.ucsd.edu
- X-Newsreader: Tin 1.1 PL3
-
- mikc@hal.gnu.ai.mit.edu (Mike Coughlin) writes:
- : In article <41910@skye.dcs.ed.ac.uk> gvw@epcc.ed.ac.uk (Greg Wilson) writes:
- : > I regularly lambast physicists for using "old" languages like Fortran,
- : > so I feel I should set a good example for students by using some
- : > modern ideas myself)
- : As a user of Fortran, I must protest the idea that Fortran is an
- : old language. Since structured programming concepts were introduced
- : into Fortran many years ago, there are no "modern" programming ideas
- : that are need in Fortran to accomplish its purpose of scientific
- : and mathematical computation.
-
- I disagree. There are some "modern" programming ideas that are necessary
- for scientific and mathematical computation, such as all the various types
- of 'tree data structures' and such things requiring dynamically allocated
- objects and some sort of pointer notion,
-
- : The mathematical notation used by
- : physicists is an old language and they don't have much interest in
- : changing it.
-
- and even worse, there are alot of useful mathematical notions that should be
- in a mathematical programming language that have been defined away as
- uninteresting by the "modern" computer science community. The surging
- importance of parallel computing (for which the primary users are
- scientists) might perhaps reverse the trend, but I'm not confident.
-
- Examples:
-
- 1) One that I'd appreciate (this has been hashed over many times) is
- array index notation. A(i,j) = B(i,k)*C(k,j). I do NOT think that
- matrix and vector computations are best done by "overloading operators"
- with a "class library".
-
- 2) Mathematicians (like the infamous Prof. Rubin) would probably enjoy
- fast multiprecision arithmetic, though this wouldn't be as useful for
- scientists.
-
- I'm sure that there are other examples, that I can't quite think of
- right now.
-
- Notation for expressing standard mathematical computations
- (I'm not talking about logic or set theory or other esoterica like that)
- hasn't been substantially improved since original Fortran.
-
- : So why should they want to switch from Fortran? Perhaps
- : if other computing languages included things like exponential operations
- : ( ** ), and complex number arithmetic there would be more reason to use
- : something besides Fortran.
-
- Agreed. It's unconscionable to make things *worse* than 1959 Fortran.
-
- : The best new idea that physicists (and every other programmer) should
- : use is extensive commenting so human beings can read source listings.
-
- I think it's the clarity of thought that makes the real difference.
-
- : That would be the greatest advance since digital computers were invented.
- : Of course if you want a language that can be modified to include any
- : new idea that you want to try, Forth is a very good choice. And I've
- : heard that Lisp and Scheme are good for that too.
-
- --
- -Matt Kennel mbk@inls1.ucsd.edu
- -Institute for Nonlinear Science, University of California, San Diego
- -*** AD: Archive for nonlinear dynamics papers & programs: FTP to
- -*** lyapunov.ucsd.edu, username "anonymous".
-