Looking for a few good programmers

[pan@ATHENA.PANGARO.dialnet.symbolics.com (Paul Pangaro)]

I shortly expect begin to a contract for the Navy for which I am looking
for both (a) advice and (b) someone to do conversion/coding.  The
situation is as follows (I will be brief but want to give a sense of it
and I ask your patience; I expect to start a more detailed dialogue with
anyone interested in continuing the conversation):

The Navy has already engaged another contractor to build a kind of
adventure game to teach electronics. It will involve mouse input,
complex graphics, animation, simulation of circuits, etc. The other
contractor is already developing all user interface code. 

One component of that game would be a place for the trainee to go when
he/she has questions, is confused, can't complete a problem correctly,
etc. The Navy is interested to our approach to training based on
individual conceptual style as embodied in THOUGHTSTICKER, a large
system written in LISP on the Symbolics (flavors, old windows,
constraint frames, etc.). You can think of this as a hypermedia system
with smarts about learning and the individual in front of the screen.
Because the i/o is done elsewhere, and only delivery of training is
involved on the PC, I estimate (roughly) that about 100K bytes of source
code need to be redesigned (slimmed down, features taken out that are
not relevant to this application, etc.), re-coded on the Symbolics (for
full testing in the context of CLOS and all the courseware authoring and
delivery interface code that is already running on the Symbolics, etc.).
Then, when all is (somewhat) well, "ported" to an MS-DOS -386 (yes, my
friends, the reality of the practical world has finally forced my hand).
That machine will have no more than 4 MB of RAM (currently an absolute
constraint!), running the adventure game and all the user i/o,
mouse/menu handling, graphics, simulations, etc., written in C++.  In
practice, therefore, somewhat less than 4 MB is available (swapping
notwithstanding).  What needs to be transferred to the -386 are
basically some well understood dbase/object definitions, manipulations
on those (make object, modify slots in particular ways, establish nested
pointers, etc.) and various heuristics that operate on the objects that
calculate a measure of cognitive learning style for the individual
trainee.

The possible ways to go are various (keep in LISP (use CLOE or IBUKI
or...), use LISP translator, re-code from scratch in C++). I will not be
surprised if this last idea will be the one we use, because machine
resources are slim; however, money resources are also constrained and so
"ease and cost efficiency" is a concern at the start. Hence the "best
way" is not at all clear. I am looking for individuals who have direct
experience in these issues who are willing to chat (on e-mail or my
phone bill) to discuss the options. Ultimately I want to find someone
who would actually do the coding for both the Symbolics testbed and the
C++ stuff, based on detailed designs.

All aid much appreciated. An e-mail discussion would benefit the net and
I encourage that but any means of contact encouraged. Thank you.

Paul Pangaro, President 
PANGARO Incorporated
800 3rd Street  NE
Washington, DC 20002   USA
Voice: 202-547-7775
FAX  : 202-547-6910
pan%athena.pangaro.dialnet.symbolics.com@riverside.scrc.symbolics.com