home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!tcsi.com!iat.holonet.net!news.cerf.net!nic.cerf.net!duncan
- From: duncan@nic.cerf.net (Ray Duncan)
- Newsgroups: comp.lang.forth
- Subject: Re: Forth's Adaptability
- Date: 27 Jan 1993 02:23:41 GMT
- Organization: CERFnet Dial n' CERF Customer Group
- Lines: 37
- Message-ID: <1k4rndINN9pn@news.cerf.net>
- References: <BEVAN.93Jan22131846@panda.cs.man.ac.uk> <1jqe2lINNiv8@news.cerf.net> <BEVAN.93Jan26102555@tiger.cs.man.ac.uk>
- NNTP-Posting-Host: nic.cerf.net
-
- In article <BEVAN.93Jan26102555@tiger.cs.man.ac.uk> bevan@cs.man.ac.uk (Stephen J Bevan) writes:
- >
- >If you write embedded code, or are otherwise not worried about
- >portability you are free to use whatever features of your FORTH system
- >you want. However, if you are worried about portability, using said
- >extensions ties you to that system until (if ever) those extensions
- >become standard or a defacto standard.
- >
-
- Sure, the choice is: try and pick a stable vendor and then use
- their extensions (knowing, that if worst comes to worst, the
- functionality across vendors is pretty much the same even if
- the word sets are somewhat different), reinvent the wheel from
- scratch yourself, or just do without such niceties as memory
- management, file interfaces, and floating point until those things
- are standardized.
-
- One point that has gotten lost in your discussion is that
- language standards tend to codify common practice. They don't
- (usually) just invent things out of thin air. If you pick a
- vendor or implementation that is widely used, your chances of
- getting badly burned by a standards process are not high -
- either the standard will look a lot like the implementation that
- you're already coding to, or your vendor will have the resources
- to migrate to the new standard and supply you with transitional
- aids or compatibility layers.
-
- For example, the functional mapping between the ANSI floating
- point, memory management, and file extensions and our existing
- systems is very close. The stack effects are different here and
- there, but most of the differences can be taken care of with
- a few lines of code layered on the existing systems. So no one
- who really cares about ANSI compatibility is going to suffer
- because they picked an LMI Forth system. And they will have had
- access to these capabilities for a decade or so before X3J14
- existed.
-
-