davidson (06/23/82)
Just as non-Euclidean geometry did not replace Euclidean geometry, new computer lanuages (for example SmallTalk) are unlikely to completely replace many old ones, for example LISP and APL, because those languages are tightly based around a particular, mathematically based model of computation. On the other hand, I dare say that hodge podges like FORTRAN, COBOL, C, ALGOL, Pascal, Ada, Modula-2, etc., no matter how useful they may be or have been will not survive indefinitely. Even when very high level languages (in which you concentrate on the what, not the how) become the rule rather than the exception, I suspect that much of what we consider essential (and much that we consider trivial) of today's computer science will survive. However, I think that this kind of mathematical nature is much more apparent at the software level. There is much less interest in inefficient (but educational) hardware, than is the case with software. Greg Davidson