Path: utzoo!attcan!uunet!hsi!stpstn!cox
From: cox@stpstn.UUCP (Brad Cox)
Newsgroups: comp.object
Subject: Re: code blocks (aka closures)
Message-ID: <5338@stpstn.UUCP>
Date: 6 Jul 90 21:10:32 GMT
References: <12396@june.cs.washington.edu> <1112@carol.fwi.uva.nl> <5305@stpstn.UUCP> <11002@alice.UUCP>
Reply-To: cox@stpstn.UUCP (Brad Cox)
Organization: Stepstone
Lines: 30

[Apologies if this was transmitted twice. We've been having networking
problems and I'm not sure if my first try succeeded].

In article <11002@alice.UUCP> ark@alice.UUCP (Andrew Koenig) writes:
:There are three problems in extending a language like C to
:support closures (Smalltalk blocks, Lisp lambda expressions
:in Lisps with lexical scoping, etc).

:I think we can all forget about translating this kind of thing to
:C, at least for the time being.

The article, "Representing Control in the Presence of First-class 
Continuations, (Hieb et all; Proceedings of the Sigplan90 Conference 
on Programming Language Design and Implementation, November 1990)
makes some fascinating claims that you should be aware of, roughly
to the effect that a segmented stack architecture could support
full-scale continuations in C without (any? much?) degradation over
C's current approach.

Although I'm keenly aware of the difficulties of pulling off what 
would probably involve a new code generation approach underneath C, 
as opposed to building new superstructures on top. But some of the 
higher-level superstructures, higher even than C++ (gate/block level) 
or Objective-C (chip level), particularly card-level systems like 
Fabrik, Metaphor, or LabView, may not be possible to do in fully robust 
form without doing something new at the foundation level.
-- 

Brad Cox; cox@stepstone.com; CI$ 71230,647; 203 426 1875
The Stepstone Corporation; 75 Glen Road; Sandy Hook CT 06482