Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!zaphod.mps.ohio-state.edu!unix.cis.pitt.edu!dwpst
From: dwpst@unix.cis.pitt.edu (Douglas W Philips)
Newsgroups: comp.lang.forth
Subject: Re: Object-oriented Forth
Message-ID: <36068@unix.cis.pitt.edu>
Date: 7 Sep 90 21:38:49 GMT
References: <29113@nigel.ee.udel.edu> <1683.UUL1.3#5129@willett.pgh.pa.us> <29762@nigel.ee.udel.edu>
Reply-To: dwpst@unix.cis.pitt.edu (Douglas W Philips)
Organization: Dislocated user association
Lines: 24

In article <29762@nigel.ee.udel.edu> carroll@udel.edu (Mark Carroll <MC>) writes:
>
>In general, I'm not opposed to the class-as-object model, but for a
>Forth system, it just doesn't seem appropriate. If you're trying to build
>a large-scale pure object-oriented system, classes-as-objects make very
>good sense. But it seems to me that for Forth, we should be looking for the
>smallest, most flexible, least restrictive model. The prototype model
>can be implemented more simply, with less code than class-as-object.

There is definitely a lot of machinery (or a non-trivial amount) to get
the full class-as-object model working.  I wonder just how much Forth
code it would be though.

>quite remember... I'll send it to you as soon as I remember. The
>papers have been published in proceedings from several sigplans. The
>best starting point is probably "Self: The Power of Simplicity", which
>is in the Sigplan proceedings of either 1986 or 1987.

I hope you'll post it so that other interested parties can know too.

-Doug

---
Preferred: dwp@willett.pgh.pa.us   Acceptable:  uunet!willett!dwp