Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Open source doesn't work

Author: Robert Hyatt

Date: 08:52:08 11/08/01

Go up one level in this thread


On November 08, 2001 at 09:52:47, Georg v. Zimmermann wrote:

>At least not for me.
>
>I write this as a hint for people starting on a chess program or any other hobby
>program, or planing to release their program.
>
>If you thinking about releasing it under GPL, "just say no" !


Don't be _too_ pessimistic.

First, what was _the_ reason you released the source?  My reason was not
based on getting a lot of suggestions, it was a "pay-back" because others
had answered lots of questions from me (Slate, Thompson, Schwartz (Fred,
from the Chaos program), greenblatt, Kozdrowicki (Coko program) and so
forth.  And I found myself answering the _same_ questions I had asked them
20 years ago.  It seemed to be a good idea to release my source, to (a) slow
down the repetition of questions and (b) give newcomers a chance to see a
real chess engine, that is not a "dumbed-down" version.

If you release the source because you hope to gain more than you lose, then
that is a gamble that might (or might not) pay off.  If you release it because
you want to help others, you win immediately.

I have gotten many source changes to crafty from others.  Some worked well.
Some needed more work but eventually ended up in the released source, some
simply didn't work at all.  There have been periods where I have been flooded
with suggestions.  And then a year here and there where I received none.  But
since I didn't release it hoping to get a lot of suggestions, I've not been
disappointed...



>
>The last chess program which I worked on and for which I released source was my
>crazyhouse program, "Sunsetter". I had the special permission of the original
>author of the program it is based on, "Deep Bug", to distribute it without code
>at first. After polishing the code I released it a few month ago in the hope to
>get some feedback or hints on low level speed improvements or programming
>technique , things I am very bad at. No suggestion ever.
>
>You might say "Well what kind of program is that anyway, probably some 2100elo
>crafty clone". Well it isn't. It has toped both ICC and Fics best lists in its
>category a while ago, if only for a couple of days.
>
>You might say "Well, who cares about this variant anway". True.
>
>So another chess related program I released under the GPL is called "Thief". It
>is a chess client for playing on ICC, Fics, ... Like winboard, but with some
>more features in some areas but no offline support for example.
>It is used for more than 10% of the games on Fics. I get around 1-7 feature or
>bug fix suggestion per week, but never *ever* has anyone apart from the original
>programmers comtributed anything, to code grafics or whatever. (Note that
>someone else did 90% of the work on this program, but he handed it over to me
>later. I am no fan of re-inventing the wheel. Others seem to be though.)
>
>You might say "Well, few people are playing chess online, and there are already
>many online chess clients". Partly true.
>
>So another GPL'd chess program I used to work on and for which source was
>available and still is on request is Zzzzz. The original program is a weak chess
>program written in Delphi. I rewrote it to play automatic lectures, and to be
>able to play without king (the one stalemated loses). The idea was that chess
>rules are complex. And its easier for people to learn by starting to play right
>away, for example only after having learned how the pawns move.
>I also made the program very easy to operate and created a way for people to
>contribute "interactive lessons". Again, some people were so kind to tell me
>they or their son/daughter liked the program. But no contribution. Ever.
>
>Don't get me wrong. I am not saying the open source idea doesn't ever work. It
>might work for big big projects like linux, or for programs like crafty (but
>here I already doubt, how many of the code improvements per version are *not*
>from Dr.Hyatt?) but else - forget it.
>
>Next I will probably try to start on a Go program. If someone wants to try that
>as a group, that would be great fun. But making it open source - no way.
>
>Georg



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.