Plans for clisp, and pcl vs clisp clos

[Marcus Daniels <marcus@sysc.pdx.edu>]

>>>>> "Raymond" == Raymond Toy <toy@rtp.ericsson.se> writes:

Raymond> Not that Bruno as moved on to bigger and better things (thank
Raymond> you very much to Bruno et. al. for clisp!), I'm curious about
Raymond> the future plans for clisp.  Is it considered "feature
Raymond> complete" now or will more things be done for clisp?

I have a code generation idea that may or may not turn
out to be generally useful.  The initial application will be for
genetic programming.. sort of a threading post-compiler.  There
is a package (still in development by Dawson Engler) called VCODE to generate
native code from virtual byte codes.  

I'm working on a GNU BFD based dynamic linker in my free time. I'd
like to have portable dynamic linking support for CLISP.

Raymond> Also, I have a question about clisp's builtin clos system
Raymond> versus pcl.  What is the real difference between them?  Which
Raymond> one is closer to CLTL2?  Which would you prefer?

The builtin CLOS system lacks some of the meta object protocol
features.  PCL is complete, but big.  See chapter 28 of the impnotes
for the lowdown vs. CLtL2.

Someday, I'd like to be able to load Objective C classes and use ObjC
methods like any other generic function.  And it seems to me the ideal
way to do this is with MOP.  So yes, it is conceivable that there
might be motivation (for me) to make some extensions to the builtin
CLOS.