Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!wuarchive!cs.utexas.edu!uunet!willett!ForthNet
From: ForthNet@willett.UUCP (ForthNet articles from GEnie)
Newsgroups: comp.lang.forth
Subject: Category 2,  Topic 8
Message-ID: <139.UUL1.3#5129@willett.UUCP>
Date: 5 Jan 90 01:27:08 GMT
Organization: Latest Link in ForthNet Chain
Lines: 31

Category 2,  Topic 8
Message 31        Sat Dec 23, 1989
NMORGENSTERN [Leonard]       at 14:16 EST

  To: Silver, Rutgers University

  >I go into cold sweats thinking about playing directly with the
  >underlying machine's guts, do you Forthers not have terrible
  >problems switching between architectures?

 Yes, we get cold sweats, a lot. I also got them the first time I drove a car
by myself. Actually, the real danger begins when you start feeling confident.

 Switching between architectures, never simple, is actually easier in Forth
than in many other languages. To illustrate, suppose that you want to use the
obscure MS-DOS call "Redirect Device" which is Int 21H Function 5H Subfunction
03H. Simple enough: write a CODE definition, and give it a suitable name, say
REDIRECT. If you port the application to Unix, and Unix has an identical or
almost identical function, you would write the word REDIRECT differently, of
course, but the action and the name would be the same. The problem arises when
the Unix version differs significantly from the MS-DOS version. Then you have
to make a decision: keep the action as it was, gaining portability, but
possibly losing efficiency under the new system, or rewrite it. And if the
latter, what should it look like? Not an easy choice of course, but once you
have made it, writing the new word in Forth is straightforward. The same
principle holds for microprocessor architectures, printers & other
peripherals, etc. The problem lies not in Forth, but in the incompatibility of
the 2 systems.
-----
This message came from GEnie via willett through a semi-automated program.
Report problems to: 'uunet!willett!dwp' or 'willett!dwp@gateway.sei.cmu.edu'