Newsgroups: comp.lang.dylan
Path: cantaloupe.srv.cs.cmu.edu!rochester!udel!gatech!howland.reston.ans.net!math.ohio-state.edu!jussieu.fr!news-rocq.inria.fr!news2.EUnet.fr!news.fnet.fr!ilog!news
From: davis@ilog.fr (Harley Davis)
Subject: Re: What's a generic function?
In-Reply-To: haahr@netcom.com's message of Tue, 18 Apr 1995 13:24:19 GMT
Message-ID: <DAVIS.95Apr18161305@passy.ilog.fr>
Lines: 32
Sender: news@ilog.fr
Nntp-Posting-Host: passy
Organization: Ilog SA, Gentilly, France
References: <3mnrf2$288r@acs6.acs.ucalgary.ca> <3mq5av$pd5@lll-winken.llnl.gov>
	<3ms6ip$4rf@stc06.ctd.ornl.gov>
	<HAAHR.95Apr18142420@netcom4.netcom.com>
Date: 18 Apr 1995 14:13:05 GMT


In article <HAAHR.95Apr18142420@netcom4.netcom.com> haahr@netcom.com (Paul Haahr) writes:

   > I imagine that some sort of call makes this at runtime, but is there
   > a way to have a "generic" call be made at compile time?

   Sealing, sketched out in the ``controlling dynamism'' chapter of the
   DIRM, is intended to allow implementations to move generic function
   dispatch to compile-time, under the author's control.

   The thing to note is that it is sometimes correct to move dispatching to
   compile-time, but sometimes that would be wrong.  Sealing in Dylan has
   richer (but more complex) semantics than C++'s distinction between
   virtual and non-virtual member function, but, to first order, they serve
   the same goal:  to control which dispatches you can do at compile-time.
   (A related aspect is which operations can be overridden.)

I am astounded by the number of people who seem to think that it is
somehow to Dylan's advantage that programmers be unable to reliably
and portably predict the approximate cost of function calls in the
language.

-- Harley Davis
-- 

------------------------------------------------------------------------------
Harley Davis                            net: davis@ilog.fr
ILOG S.A.                               tel: +33 1 46 63 66 66
2 Avenue Gallini, BP 85                fax: +33 1 46 63 15 82
94253 Gentilly Cedex, France            url: http://www.ilog.com/

           Ilog Talk information: info@ilog.com
