I don't know if somebody may be interested, but Fred Rivard and me wrote in 1991 an interpreter of Self 1.0 in Smalltalk 2.5 (Ultimardrev
Self). It contains a parser of the full Self-syntax, that was able to reload most of the system files, for generating the initial Self world. The sources may be available.
While it never hurts to look at other people's code, it is main in their *ideas* that I am interested in.
Then starting from this interpreter, Pierre Cointe, Jacques Malenfant, Christophe Dony and me have extended its virtual machine to decompose message sending according to the lookupOapply model, supported by a meta-object protocol. This is an alternative to limited mirrors for reflecting upon Self.
Could you elaborate a little on that? I have read all I can get my hands on about reflection, but haven't heard about lookup/apply. I found a lot of good stuff via ftp to p300.cpl.uiuc.edu in the /pub/washington directory.
Actually, we are defining our own full-reflective plateform, Moostrap, inspired by some Self constructions, and we would be very interested by any experience with reflection in Self.
I am also trying to extend the reflection in Self in my Merlin Project. You can read a little about it in the Web as
<a href=http://lsi.usp.br/~jecel/merlin.html> Merlin </a>
Unfortunately, the part about reflection is among those that need to be greatly expanded. I based most of my ideas on the Apertos OS developed at Sony ( see ftp://scslwide.sony.co.jp ).
=================== //\ Merlin Computers Jecel Assumpcao Jr / /\ \ + LSI - Universidade de Sao Paulo firstname.lastname@example.org '-/-----' + Pontificia Universidade Catolica =================== '------' Sao Paulo - Brazil