Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Bionic v Crafty - a possible solution

Author: Bruce Moreland

Date: 00:42:18 01/26/99

Go up one level in this thread



On January 26, 1999 at 01:24:11, James Long wrote:

>I think it is possible to view another's source, learn from it,
>and continue the creative process.  Some degree of discipline is
>involved.  It is not so different from reading a technical report
>and implementing the theory.  In some ways, your program is
>still based on the works of others.  Maybe this is a hard pill
>to swallow for those that want to claim sole ownership, but
>it's true.

One reason that computer chess is interesting is that the products are not only
distributed, either downloaded for free or sold, but that the products compete
in semi-regular international tournaments.

Anyone is free to devise whatever relationships they want if the point is only
distribution (sale).  I can make some arrangement with another programmer, and
include bits of his code in my own thing, and that's no problem.

The competition aspect creates the problems.  Each program needs to have the
name of an author or group of authors attached to it, in order to determine
eligibility for competition.

If I write two programs, there is an extremely good chance, in most cases 100%,
that I will not be allowed to compete with both programs.  And if I am deemed to
be co-author of several programs, some of the programs may be disqualified from
competition.

This makes sense for a lot of reasons.  A main criterion for selection of
entries for competition is strength.  A strong author might write several strong
programs, and get accepted several times.  This would squeeze out the weaker
programs, which would deprive their authors of experience, and would also give
the stronger author a greater chance of winning the competition, potentially
resulting in increased commercial success and further compounding of the problem
next time.

It's not a matter of reinventing the wheel.  We all read the literature, the
literature predates the competition, and everyone has equal access to it.  We
don't reinvent the wheel, we make use of that research that we find out about.
But if you take code directly, and I'm talking about significant chunks of code,
and not just a snip here and there, you start to muddy up this authorship issue.
 Especially when you you start from another program as a *base*, because it is
my opinion that in cases like Crafty, a lot of the strength is in the parts of
the program that are not sexy and would not be rewritten.  People can talk about
tweaking extensions or rewriting the eval function, but just the bitmap move
generation/execution stuff, the hash table, the parts of the search that allow
it to run parallel, the book learning, etc., are a heck of a lot of work, and if
you put piece-square tables, and rudimentary pawn and king safety evaluations,
on top of that you'd have a pretty strong program.

Bob is clearly a co-author of anything that starts from Crafty.  It may be
possible to take this same viewpoint with regard to Gnuchess and its original
authors, but Gnuchess isn't as strong, so there is less chance of pushing out
weaker programs, but additionally it is fairly simple, so I can much more easily
believe that someone can put in enough work to drown out the work put into that
program by the original authors, especially given that anyone who started with
Gnuchess has been doing this for *several* years.

Bob got the Novag award, which recognizes contribution to the field of computer
chess, for Crafty and for his other internet contributions.  I fully understand
and recognize this, in fact I nominated Bob for the award.

The point of that award was contribution to computer chess.  I don't want that
contribution to turn into a detriment, which is what I am concerned will happen
if Bob becomes an unwilling co-author of a lot of these programs, and they are
allowed to compete as completely unique entries, and drown out the innovation
that comes from writing each new program from scratch.

bruce



This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.