Re: Porting LISP code from SYMBOLICS

[Joe Karnicky <KARNICKY@KL.SRI.COM>]

	My principal activity here at the Varian Research Center is
the initiation and development of knowledge-based application programs for
analytical instruments.  We have a collection of inferencing software
that has been developed over the last 5 years which we call VIS
(Varian Inferencing System).  The way these projects usually go is
that I work for a couple of years in cooperation with one of our
instrument divisions (e.g. NMR, HPLC) at the end of which time program
responsibility is turned over to the division.
	VIS is written in Common Lisp and makes sparing use of objects.
I find the most efficient way to work on these programs is to develop
the code in GENERA on one of our Symbolics 36**'s and later port to the target
platform.   Depending on the complexity of the program and the degree
to which the target language adheres to the Steele spec. (which goes better
and better all the time) the port is a matter of a few hours to a few 
days.
	We've ported to the following environments:
	MACHINE			OS		LISP
	-----------		---		-------
	VAX and MICROVAX	VMS		DEC CL
	MICROVAX		VMS		LUCID CL
	SUN 4/110		UNIX		SUN CL (actually LUCID)
	286 and 386		MS-DOS		GOLD HILL CL

As you might imagine, many many issues have been dealt with over the
years (speed, cost, portability, user interfaces, remote access, etc.
etc.) I'd be glad to correspond with you further on the specifics.

Joe Karnicky
<KARNICKY@SRI.STRIPE.COM> but actually at Varian in Palo Alto
(415) 4245085




-------