Re: LMI random systems

[Alfred M. Szmidt]

   On 2024-11-26 16:57, Alfred M. Szmidt wrote:
   > How do we want to handle random LMI systems?
   >
   > I just compiled ObjectLisp (some Flavor inspired thing ...); it is
   > part of "SYS:".  But I feel like our system shouldn't include such
   > things?
   >
   > Similar, Verify, LM-Prolog make sense that they live somewhere "else"?
   >
   > I'd even go as far as .. ZMail should live by it self, and even Tape.
   > But since they traditionally lived beside everything else I'll just go
   > with it.

   If you mean those systems should be separate from System, I agree! Or 
   what do you mean by "somewhere else"? Elsewhere than on the SYS host? 
   Why? (They can be on SYS without being under /sys, /l, /tree, or 
   whatever is this week's name of it.)

They would live under ~/l or whatever one uses.  But they would not be
(unless sys.translation rules ...) SYS: host (which would be ~/l/sys
or something)

   I would argue quite strongly that Zmail, Tape, and possibly perhaps 
   LM-Prolog should *definitely* be on SYS *and* under /sys. Taking Zmail 
   out would be like taking Converse out.

I'll not argue in the least against, or for it .. I'd prfer to take
them out but it is for "to much information" reasons.  I do understand
why ZMail and Converse (at least) should be part since they allow
colaboration (even if ZMAil is not very useful -- yet).

Having said that ..  LM-Prolog will not be under ~/l/sys (or whatever)
though.  LM-Prolog is under ~/l/lm-prolog (or whatever your root is
:-) and a simple lm-prolog.translation file like the following:

(fs:set-logical-pathname-host "LMP"
  :physical-host "OZ"
  :translations '(("SYSTEM;" "//lm-prolog//")
                  ("*;" "//lm-prolog//*//")))


will fix it (similar to Macsyma)