home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
ftp.cs.arizona.edu
/
ftp.cs.arizona.edu.tar
/
ftp.cs.arizona.edu
/
icon
/
newsgrp
/
group93a.txt
/
000055_icon-group-sender _Mon Jan 18 01:16:44 1993.msg
< prev
next >
Wrap
Internet Message Format
|
1993-04-21
|
2KB
Received: by cheltenham.cs.arizona.edu; Sat, 30 Jan 1993 20:39:53 MST
Path: ucbvax!dog.ee.lbl.gov!overload.lbl.gov!agate!ames!haven.umd.edu!uunet!mercury.hsi.com!mlfarm!cs.arizona.edu!icon-group
From: reid@vtopus.cs.vt.edu (Thomas F. Reid)
Newsgroups: comp.lang.icon
Subject: Re: endtab/detab
Message-Id: <9301180116.AA05090@vtopus.cs.vt.edu>
Date: 18 Jan 93 01:16:44 GMT
Lines: 32
In-Reply-To: <199301161259.AA02671@cheltenham.cs.arizona.edu> from "Ralph Griswold" at Jan 16, 93 05:59:45 am
Apparently-To: icon-group@cs.arizona.edu
Status: R
Errors-To: icon-group-errors@cs.arizona.edu
> However, once you have a large user community and widely distributed
> documentation, it's not feasible to remove something, even if it's
> little used. Worse, it's not feasible to redesign something that
> was done badly. That's one of the reasons language design is so
> hard; once you've done it, implemented it, documented it, and
> released it, you're stuck with it.
>
> Ralph E. Griswold ralph@cs.arizona.edu
> Department of Computer Science uunet!arizona!ralph
> The University of Arizona 602-621-6609 (voice)
> Tucson, AZ 85721 602-621-9618 (fax)
>
I think that a change of paradigm is needed for language design. I
believe that the language designer should be encouraged to release a new
language design and implementation every (say) 5-8 years along with a
translator to/from the new/old versions.
This way, the language designer can correct old mistakes and extend the
language "right" into new concepts without paying too much homage to the
cursed god of upward compatibility. There have been too many elegant
languages ruined by trying to extend them into areas that the original
design cannot comfortably permit. Modula-2 comes to mind.
If the two versions are fundamentally the same, then the delta between
implementations and the to/from translators should be straightforward.
It is extra work for the language designer, but it has the reward of
allowing the language to evolve gracefully rather than stagnate or accept
awkward extensions.
Whatcha think?
Tom