COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	.require "let.pub[let,jmc]" source
C00006 ENDMK
C⊗;
.require "let.pub[let,jmc]" source;
∂HOM Mr. Chuck Williams↓Inference Corp.
↓5300 West Century Blvd.↓ Los Angeles, CA 90045∞

Dear Chuck:

	I have been thinking about the SMP2 project, and I have a few
suggestions.

	1. At present there is not a clear idea of what features would
make SMP2 more valuable to users of symbolic computation than
SMP or MACSYMA.  It may well turn out that after the Government
money is used up, it won't be in Inference's interest to develop
SMP2 further and market it.

	2. Rich Schroeppel suggested, and I agree, that more information
about what users are doing and what limitations they find are
needed.  I suggest he be asked to survey this.  The simplest kind
of survey is to place telephone calls and ask.  Visits to a few
of the more knowledgable of those contacted may be desirable.

	3. My guess is that many respondents will grumble that SMP
balks at problems that are too big in various ways and will most
value improved algorithms that fix this.  This won't be the complete
story, because users often don't think about genuinely new features.

	Let me also mention a few suggestions that I made at the
first Board meeting and now found in my files.

	1. A %2Physics Today%1 article by Wolfram or other about
symbolic calculation in physics would be good.

	2. A package of sample symbolic computation applications
could be compiled by collecting SIGSAM articles.  Inference
should acquire a complete file of SIGSAM.  If necessary, the
issues should be xeroxed at some University library.

	3. The ad should have some real examples as the Macsyma
ad does.  One theme for an ad might be "Well, I'd work for you
if you had SMP".

	4. A version of SMP for micro-computers too small for it
to be more than a toy is still worthwhile.  Many users will play
with it and use it to teach themselves integration or other aspects
of symbolic computation.  Since SMP is in C, this shouldn't be
difficult.

.reg

cc: Rich Schroeppel, Alex Jacobsen