🚀 go-pugleaf

Seeing this portion of a fri.26.jun.1998 post by mentifex@scn.org...

   http://www.scn.org/~mentifex/ Project Mentifex promotes DEC Alpha
   as used in the Cray T(orus)-3D and other 64-bit architectures for:
   1. the memory space made unfillably large as compared to 32-bit;
>> 2. the possibility of an 8 x 8 visual array within CPU registers;
   3. built-in INSTANT accessibility of a mind core of 64 concepts.

...Andy Glew "glew@hf.intel.com" on mon.29.jun.1998 posed a question:

> What are the advantages of an "8x8 visual array",
> versus the 4x4 transformation matrix typically used
> in computer graphics texts?

First off, the Web has a growing cluster of robot vision sites:
http://dv.op.dlr.de/FF-DR/dr_rs/VISION/vislinks.html Robot Vision Links;
http://members.aol.com/krobin8082/vision/page1.html Vision web;
http://www.newtonlabs.com/ The Cognachrome Color Vision System;
http://www.wspc.com/books/compsci/1960.html "ACTIVE ROBOT VISION".
<PRE>
  /^^^^^^^^^^^\ The Architecture of a Robot Brain /^^^^^^^^^^^\
 /visual memory\           semantic ________     /  auditory   \
|      /--------|-------\  memory  / syntax \   |episodic memory|
|      |  recog-|nition |          \________/---|-------------\ |
|   ___|___     |       |  flush-vector|        |    _______  | |
|  /image  \    |     __|___        ___V___     |   /stored \ | |
| / percept \   |    /deep  \------/lexical\----|--/ phonemes\| |
| \ engrams /---|---/concepts\----/concepts \---|--\ of words/  |
|  \_______/    |   \________/    \_________/   |   \_______/ </PRE>

Although visionaries rightfully insist that there is an "intelligence"
within the human visual system, the above depicted Theory of Mind
DOES NOT CARE how visual recognitions are made, but only that a
unitary recognition signal is quickly flashed into the mind core.

Therefore the advantage of an "8x8=64" visual array is that it
BUYS US TIME while the immense and stupendous difficulties of robot
vision are worked out by the geniuses for whom we have been trolling.

A 64-bit array can recognize (how many? 2 to the 32nd power?) pre-set
images--not four billion "random" images, but several hundred non-fluid
COGNITIVELY SIGNIFICANT RASTER PATTERNS without having to analyze them.
The 64 bits can even be a table-top world for an AI alife to inhabit.