Path: utzoo!utgpu!news-server.csri.toronto.edu!bonnie.concordia.ca!uunet!spool.mu.edu!sol.ctr.columbia.edu!ira.uka.de!smurf!artcom0!hb.maus.de!ms.maus.de!Kai_Henningsen
From: Kai_Henningsen@ms.maus.de (Kai Henningsen)
Newsgroups: comp.lang.eiffel
Subject: Re: Functions without side eff
Message-ID: <17726@ms.maus.de>
Date: 18 Jun 91 00:42:00 GMT
Distribution: world,comp
Organization: Maus Mailbox Netz - UUCP-Gateway Bremen
Lines: 28


William F Ogden ogden @ seal.cis.ohio-state.edu schrieb am Do 06.06.1991,
23:05

WF>As it turns out, referencs semantics are fine -- provided you only allow
WF>_one_ reference to each object. (With this constraint, you effectively
WF>have value semantics, of course.) Now the problem, rephrased in these

Has anybody ever considered mirroring this as a language construct? Probably
not too clear :-)

What I mean is, differentiate between "primary" and "secondary" references,
with the semantic that secondary references allow only read access to whatever
they reference - of course that would have to be enforced somehow.
Then, you might constrain the system to only one reference at a time, more to
be created via clone operations.

Hmmm. As I think of it, you could even do what modern implementations of fork()
(another expensive directive) does, and implement copy-on-write techniques. You
would have to keep book on whether an object reference is currently modifiable
or copy-on-write, of course.

That might even be *more* effective, as you could clone only as much as
absolutely needed ...
--
Kai Henningsen  Internet: kh@ms.maus.de
Muenster         UUCP: any_backbone_that_knows_domains!ms.maus.de!kh
Germany         Fido: kh%maus ms, 2:242/2.6 or Kai Henningsen, 2:242/2.244