Re: Art of MOP

[bill@cambridge.apple.com (Bill St. Clair)]

>Questions about MCL and MOP:
>1. Is there an available document describing MCL p3 and final implementations
>of MOP?

I have a text document describing the subset of the introspective MOP
that MCL defines. I have not included it in this message as it is
a little over 20K long. If you want a copy, just ask.

MCL 2.0 does not contain a complete MOP, only a subset of the
introspective MOP.

>2. Does PCL run on MCL? Any tips for getting it running?

It will not run without work. I remember someone trying to bring
it up over a year ago, but don't remember his name, nor do I remember
if he succeeded. Bringing up PCL in 2.0 final will be difficult because
MCL no longer includes built-in support for funcallable instances. We
do our generic functions a different way.

>3. Does Art of MOP represent current MOP design, or has the "standard" changed
>since that publication?

I hesitate to say too much about this. The AMOP is the only sensible
description of a MOP that I know about. I found the earlier MOP writeups
to be reasonable descriptions of the PCL implementation, but they would
have been difficult to use as a basis for a new implementation. If I were
to write a MOP for MCL today, I would base it on The Art of the MOP.
I remember seeing a writeup from Gregor about more recent work at PARC in
meta-level languages, but don't remember the details.