🚀 go-pugleaf

Hello,

I bought an old book: "Chess skill in man and machine".

In this book there was a chapter 10:

"Using Chess knowledge to reduce search"

This handled about a chessprogramm: PARADISE, 
which solved chessproblems processing only few hundreds of possibilities,
because it played with a plan and was using a lot of chessknowledge to guide 
the search.

I'm very interested in PARADISE. Can someone tell me more about it?

What has become of the programm?

What are the conclusions, except that it is difficult to implement
chessknowledge?

And technical questions:
Can someone send me all Knowledge Sources, Concepts etc, that
Paradise used, or any other programm that looked like it?

I'm also interested in the hardware it was running on:
a DEC KL-10. How fast was this machine, compared to the current
hardware i have: an Intel 486dxII.

Yes, you gambled right, i'm beginning to write a second chessprogramm
that will look like Paradise( my first programm, were i still work
on is just another brute force programm; something like the King was
during it first years of development ), just because alpha-beta is 
to slow for a 486, if you want to solve problems that need over 
about 7,8, or more than or equal to 9 ply, or if you want an 
evaluator that evaluates a position 
just like a human does (and not giving points away for 
advantages/disadvantages, without knowing what advantage is most 
important).

Greetings, 
Vincent Diepeveen.

Everybody who wants to tell something, ask something, or just chat:
please write to:
vdiepeve@cs.ruu.nl
or in this area.







--
                +--------------------------------------+
                ||  email  : vdiepeve@cs.ruu.nl       ||
                ||  fidonet: 2:280/206.23             ||
                +======================================+