home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
ftp.cs.arizona.edu
/
ftp.cs.arizona.edu.tar
/
ftp.cs.arizona.edu
/
icon
/
newsgrp
/
group94b.txt
/
000116_icon-group-sender _Mon Nov 21 01:07:51 1994.msg
< prev
next >
Wrap
Internet Message Format
|
1995-02-09
|
1KB
Received: by cheltenham.cs.arizona.edu; Mon, 21 Nov 1994 00:08:04 MST
Date: Mon, 21 Nov 1994 01:07:51 +0600
From: jeffery@runner.jpl.utsa.edu (Clinton L. Jeffery)
Message-Id: <9411210707.AA09975@runner.utsa.edu>
To: ruiter@ruls41.fsw.LeidenUniv.nl
Cc: icon-group@cs.arizona.edu
In-Reply-To: <3aofof$f5m@highway.LeidenUniv.nl> (ruiter@ruls41.fsw.LeidenUniv.nl)
Subject: Re: optional typing in Icon (longish)
Content-Length: 973
Errors-To: icon-group-errors@cs.arizona.edu
> Jan Peter de Ruiter suggests we add optional type declarations to Icon...
The Icon implementation book contains much useful material based on the Icon
version 6.0 translator/interpreter system (icont), but it does not report on
the last 7-8 years' of work on Icon. The Icon compiler (iconc) does type
inferencing and determines a precise type for over 90% of the uses of
variables. It produces code that skips run-time checks when they are not
needed. Your suggestion would speed up the Icon interpreter somewhat, but
would not help the Icon compiler much.
I think optional type declarations are a good idea anyhow, because in larger
(multiperson) projects, they would serve as documentation and as an
assertion that the Icon compiler could verify during type inferencing.
Unfortunately, there are presently no resources to pursue this kind of
effort, so your options are: write your own, or form a group that will
support these kinds of additions to the language.