Newsgroups: comp.lang.scheme,comp.lang.scheme.scsh
Path: cantaloupe.srv.cs.cmu.edu!nntp.club.cc.cmu.edu!goldenapple.srv.cs.cmu.edu!das-news2.harvard.edu!cam-news-feed3.bbnplanet.com!cam-news-hub1.bbnplanet.com!news.bbnplanet.com!news.maxwell.syr.edu!ott.istar!istar!news.istar.net!van-bc!news.mindlink.net!nntp.portal.ca!news.bc.net!torn!kwon!watserv3.uwaterloo.ca!undergrad.math.uwaterloo.ca!not-for-mail
From: papresco@csclub.uwaterloo.ca (Paul Prescod)
Subject: Re: Ousterhout article on scripting applies to scheme
Sender: news@undergrad.math.uwaterloo.ca (news spool owner)
Message-ID: <E80wHu.p6y@undergrad.math.uwaterloo.ca>
Date: Wed, 2 Apr 1997 18:25:06 GMT
X-Newsposter: Pnews 4.0-test50 (13 Dec 96)
References: <rcybba5k9c.fsf@redwood.skiles.gatech.edu> <rcn2rprri4.fsf@redwood.skiles.gatech.edu> <E7rI6z.nzE@undergrad.math.uwaterloo.ca> <j7vd8se16hp.fsf@new-world.cs.rice.edu>
Nntp-Posting-Host: calum.csclub.uwaterloo.ca
Organization: University of Waterloo Computer Science Club
Lines: 34
Xref: glinda.oz.cs.cmu.edu comp.lang.scheme:19270 comp.lang.scheme.scsh:458

In article <j7vd8se16hp.fsf@new-world.cs.rice.edu>,
Shriram Krishnamurthi  <shriram@new-world.cs.rice.edu> wrote:
>papresco@csclub.uwaterloo.ca (Paul Prescod) writes:
>
>> >Hopefully when Guile arrives on the scene, many of these things
>> >will be eventually available for it (primarily due to the higher
>> >profile of the FSF and GNU project).
>> 
>> I hope so too. A standard FFI would go a long way to integrating these 
>> facilities without integrating actual Scheme implementations.
>
>Maybe I'm missing something here, but how is Guile going to solve this
>problem?  If you want _any_ FFI, there are several around already:
>Chez's, MzScheme's, Scheme 48's, etc.  These exist, are being used
>daily, and have had a fair amount of effort put into them.  They're
>being used to build the things that make things work (pardon, Lucent).
>
>If you believe Guile is going to announce one that will cause every
>other Scheme implementor to modify theirs to toe the Guile line, I'll
>just leave it at saying I think you're being overly optimistic (and
>that you ought to be be prepared for disappointment).  Especially if
>the Guile-ordained interface runs contrary to the research other
>implementors have done to decide their interface works "best" for the
>tasks it was designed for.

I meant that in the absence of a standard FFI, a "standard" Scheme
implementation (in the sense that GCC is the "standard" C++ 
implementation on Unix platforms) would be better than nothing.
A standard FFI would of course be better: EVEN IF IT IS SLOW. I don't
need lightening speed to tell the GUI to pop up a window or the 
OS to open a network connection and send this string across it.

 Paul Prescod

