RE: academic freedom (The Mentifex Manifesto)

["M. Edward Borasky" <znmeb@xxxxxxxxxxxxxxxx>]

----------
From: 	George A. Stathis[SMTP:hyperlex@hol.gr]
Sent: 	01 December, 1996 14:29
To: 	Jeff Fox; MISC@pisa.rockefeller.edu
Subject: 	Re: academic freedom (The Mentifex Manifesto)

I guess I'd better introduce myself. I write PROLOG code, Assembly code,
'C', etc, and focus a lot on Natural Language Understanding...
I've written theorem-provers based on George Spencer Brown's "Laws of
Form", and was truly amazed to discover the activities about MISC, and
the parallels between Laws of Form and Forth. Though I'd request some
further explanations, if at all possible, and some source code to
illustrate the parallels. If there is hesitation due to mistrust, I
can compensate by sending my own code (though please bear in mind this
may take time -I also run a small software business and time is scarce).

I've seen some other references to Forth and Laws of Form, but I can't remember where I saw them.  You should post on 'comp.lang.forth'.  There are also implementations of LISP-like and PROLOG-like constructs in Forth; try the Forth Dimensions and FORML keyword indexes available from the Forth Interest Group.

I am in communication with Dr. Nick Vainos in Crete (the "Research
Park" of companies affiliated to the university there), about theorem
proving using a variant of Brownian Logic implemented entirely in
Optical Hardware. The high level of parallelism of such Hardware
is amazing, but there are some problems. (e.g. the tendency of quartz
crystals to "forget"... )

Nick and I are close friends from the past, but we have a small
disagreement about what is best to do _now_, (before Massively
Parallel Optical Computers come into the scene): I look into silicon
processors more, and feel it may be worth working with them first.

So I discovered your list, after browsing at F21 and downloading S21...

I am also of the opinion that silicon technology has a long way to go before something else replaces it.  Without the Cold War, there really isn't anything to drive the search for higher-performance technologies.  That is, the only place where, say, optical computing would be cost-effective would be in a supercomputer for highly-classified military projects.  These days, even conventional supercomputing is struggling to find a market.


I am mostly concerned with applying Brownian Logic to Automatic
Translation. (Released a product, highly unintelligent, three years
ago, in Greece, but prepating the "next version" using Assembly-
Language-based simulations of parallelism -with Brownian Logic).

If you have hardware or software needing clients and _markets_,
I am your man in Greece: A pleasure and an honor to be your client.

And if your hardware can help in speeding-up Brown Theorem-Proving,
and is also affordable, I will build a (new-age-) altar for you! :-)

I would guess that the P21/F21 would indeed be a very low-cost high-speed engine for such a project, as long as the address space is sufficient.  As I understand it, the F21 has parallel capabilities similar to those of the Transputer, so the address space limits of a single processor would not necessarily be a problem.  I don't think the P21 has this capability, though.

One thing puzzles me, though.  I'm not sure how Brown's Laws of Form differs in any fundamental way from more conventional Boolean logic.  In the end, don't all forms of 'logic' reduce to Boolean matrix operations (which the P21/F21 ought to be blindingly fast on!)?