"Portable" Objects Package Available?

[RWK@F.ILA.Dialnet.Symbolics.COM]

Konban wa, Kurain-san

    Date: Tue, 13 Mar 90 13:11:57+0900
    From: kddlab!harl.hitachi.co.jp!mklein@uunet.uu.net

    I have a body of code that uses flavors (originally on the Explorer,
    now on the Symbolics) that I will have to port several times among
    differing LISP machines as well as probably to SUNs. I would like
    to use a portable Objects package, if such a thing exists, so I do
    not have to modify the flavors stuff every time I change machines. I hear 
    that there is a new Common Lisp Objects standard called CLOS.
    How "standard" is it? 

At this point, it's as official as anything else in Common Lisp.
It has been voted on and been accepted by the ANSI Common Lisp
committee as a core part of the language, and thus will be part
of the draft standard.  However, the draft standard will be subject
to a round of voting, too, so NOTHING is COMPLETELY officially
standard yet.

			  Is source code available that I can 
    port to different machines (especially the Symbolics)?

Yes.  You won't even have to do any porting.  There is a package
called PCL which is available from Xerox.  All you have to do is
load the tape and compile it.

  CommonLoops Coordinator
  Xerox PARC
  3333 Coyote Hill Rd.
  Palo Alto, CA 94304
(or send Arpanet mail to CommonLoops-Coordinator.pa@Xerox.arpa)

I presume they charge a small tape-copying fee for the tape, but
I don't know how much.

Or closer to home, you probably can get a copy from Nihon Symbolics.  Or
perhaps Hitachi's AI department in Yokohama may have a copy.  Try
calling group-leader Morita-san there.  If I read his meishi right, it
says Denwa (045) 824-2311 daihyoo (typical), (045) 821-4111 yakan
chokutsuu (evening direct line), (045) 824-9092 fuakkusu (fax).  That's
the office at Totska-ku Shinanoo-cho 549-6.  Near Totsuka-eki if my
memory serves.

If that doesn't work, his meishi also lists another building nearby (in
Totsuka-ku Totsuka-cho 5030 Banchi; denwa (045) 881-7161 daitaihyoo
(?very typical?).  (Too bad I can't send you the kanji).  But I think
it's the first set of numbers that's more likely.  I hope that helps.

Any copy should run on almost any Common Lisp available.  There are a
few files which are specific to specific implementations, but the release
includes those files for all the lisps it runs on.  If you do get
a copy of PCL in Japan, Xerox requests that you send mail to the above
address with your name and mailing addresses (physical and network).

(defvar *port*
        '(#+Genera               Genera
          #+Genera-Release-6     Rel-6
          #+Genera-Release-7-1   Rel-7
          #+Genera-Release-7-2   Rel-7
	  #+Genera-Release-7-3   Rel-7
          #+Genera-Release-7-1   Rel-7-1
          #+Genera-Release-7-2   Rel-7-2
	  #+Genera-Release-7-3   Rel-7-2     ;OK for now
	  #+Genera-Release-7-4   Rel-7-2     ;OK for now
	  #+imach                Ivory
	  #+Cloe-Runtime	 Cloe	     ;Symbolics IBM-PC Lisp
          #+Lucid                Lucid
          #+Xerox                Xerox
	  #+Xerox-Lyric          Xerox-Lyric
	  #+Xerox-Medley         Xerox-Medley
          #+TI                   TI
          #+(and dec vax common) Vaxlisp
          #+KCL                  KCL
          #+IBCL                 IBCL
          #+excl                 excl	     ;(Franz Allegro)
          #+:CMU                 CMU
          #+HP-HPLabs            HP-HPLabs
          #+:gclisp              gclisp
          #+pyramid              pyramid
          #+:coral               coral))     ;(Apple Macintosh)

However, many venders are about to bring out "native" version of
CLOS, which will replace the use of PCL.  For example, Symbolics
Genera 8.0 includes CLOS.  (I don't think I can legally tell you
who else is doing this yet; you'll have to wait and see.  Since
I haven't signed any non-disclosures with Hitachi, I think I can
tell you that Hitachi was planning to do so, the last I knew).

    Ideally I would like to have a chunk of Common Lisp code, including 
    the CLOS source if neccesary, that can run with few/no modifications on
    any standard Common Lisp interpreter.

I think you'll have a lot more problems with incompatibilities
with pathnames than you will with object systems.

Probably the worst part will be converting from message-passing syntax
to generic-function syntax, but I think that's worth it to make the code
easier to maintain and extend.  It's not too hard to write editor tools
to do this conversion.

    I would appreciate any suggestions people may have in this regard.
    
	    Thanks,
    
		    Mark Klein, Ph.D.
		    Hitachi Ltd.
		    Advanced Research Lab
		    Higashi Koigakubo 1-280
		    Kokubunji, Tokyo 185
    
		    mklein@harl.hitachi.co.jp
		    0426-23-1111 ext. 2822