Brian Rice's intro mail (forwarded)

[Tril]

Brian sent this message to Fare, who sent it to me.  Brian and I have had
some e-mail discussions since then, which we are now continuing on the
list.

Date: Sat, 3 Oct 1998 23:13:51 -0700
From: "RE01 Rice Brian T. EM2" <BRice@vinson.navy.mil>
To: "'rideau@ens.fr'" <rideau@ens.fr>
Subject: a TUNES-related prospect!! Please read carefully!!

Greetings.
   My name is Brian Rice.  I am a self-taught programmer and amateur
mathematician who has been using your TUNES project's documentation to
help with my intense research for the last 2 years.  I am deeply
interested in your goals with this project, although the viewpoint that
I have taken on it has given me, I believe, quite an advantage from
which we should both benefit.
   I first learned of your project from my own research efforts on the
WWW.  I can say, after about three years, that I have fully covered the
sites and issues that your TUNES documentation and reviews mention.  I
therefore believe that I could be quite a help, at least in the sense of
an adequately-educated point of view on your development.  However, that
is not entirely why I am speaking to you now.
   I am working on quite a similar project, spawned from a software
system I envisioned over 5 years ago, as a 16-year old student.  I was
mostly inspired by the goal of HAL-9000 equivalent performance and a
level of mathematical sophistication to allow a more intelligent
interface (not necessarily artificially intelligent).  However, I did
not have the kind of mathematical experience to understand what sorts of
things I was talking about.  I understood a concept of describing to the
computer information as a variation of linked lists forming the 'shape'
of the data.
   To make a long story short, I now have the practical understanding
needed to create such a system, and am currently working out the final
details of the specification for my software system's 'kernel'.  The
essential feature of the kernel is that it contain a complete reflective
system of logic (I will explain this later).  The essential primitive
action of the runtime (persistent) system is to take this ultimate
logical form and reify some aspect of the environment to create a
'sub-environment'.  This action creates a hierarchy of partial orderings
which, due to the aspect-orientation capabilities and reflection, can be
arranged by the user to make anything, including an operating system, an
API emulator, a language 'compiler'/'interpreter'/etc...., an interface
specifier (human-computer as well as computer-computer), and all the
associated applications that can be imagined.
   This system will stick to every detail of the TUNES specification and
refine and generalize it somewhat.  The 'hll' would be the reflective,
aspect-oriented logic of the kernel, since all other logical atoms would
be semantically-defined in terms of this.  The 'lll' concept is very
vague due to this point: the means by which the task-kernel,
file-system, device-drivers, and code generators would be implemented is
through semantics-based descriptions of the computer to the kernel.  In
other words, the kernel implements itself onto a computer in an
appropriate representation.  I believe that you should now have the
insight to see how 'portability' is achieved: via machine-descriptions
of the target machine and the appropriate applications of these
descriptions (including the method of installation).
   Well, I could continue, but I would really appreciate your feedback
and criticism.  Please believe that after three years that I have
considered most every issue in minute detail, though I don't wish to
seem inflexible.  However, I believe most arguments could be easily
countered by research which your project references, so please please
please (with sugar on it :) let's just start discussing some sort of
partnership (go ahead and ask all the questions you wish, too).  One
major thing which you should know is my intention to request the help of
the famous Linux kernel hackers who seem to know hardware quite well, as
well as the VHDL specifications and tech manuals for hardware
descriptions so that the rest of development after the logical kernel
will be quite trivial.

			Hope to Hear from You Soon!
				Brian Rice