[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [computer-go] future KGS Computer Go Tournaments - two sections?



This whole business of clones and similar programs is REALLY 
about AUTHORSHIP.    What makes any of this fair or unfair has to do with
the statistical unfairness of having multiple entries. 

It has nothing to do with strength difference and shouldn't.  Unless
we are screening for tournament quality for some kind of championship
tournament, it's not about how much stronger one program is that is
the criteria for how "different" it is.  It shouldn't be playing style
either.  Even though one GnuGo derivative may play like GnuGo, I don't
think THAT has anything to do with why it shouldn't be allowed.

For instance, I have more than one Go program, and they use different
algorithms.  They don't just use different algorithms, they didn't
even start life as the same piece of code.  I wrote ogo/botnoid from
scratch as a completely different program.  I have brute force
searchers, highly selective searchers, monte carlo based ones and even
the monte carlo programs are not much alike, some use population based
learning algorithms, others use simple statistical selection
mechanisms and botnoid does a more sophisticated analysis of move
SEQUENCES that is original and unpublished.

But why can't I enter several of these programs in competition?  The
similarity between any 2 of these is FAR less than the similarity
between SlugGo and GnuGo.  

The answer is authorship.  As David and others have pointed out, it's
grossly unfair to let me or anyone else have several programs
represented in a single tournamnt.

But now many people are proposing that if SOMEONE ELSE MAKES THE
CHANGES, IT IS SUDDENLY OK.  Are you serious?  All I have to do is
give the source code to someone else, let them make changes and now
they can be called original authors and I can have 2 entries?

So I can't play 2 completely different program (that I wrote) even
though it would provide far more diversity to the registry of a
tournament, but someone could take a different version of the SAME
program and many of you would be happy?

If someone takes GnuGo and makes some changes to it,  what you have
is a modified version of GnuGo,  not a different program.  

If you look at the GnuGo project, you will see that GnuGo has many
different authors.   In FACT, I am technically one of them because 
I contributed a little GUI.   

So why can't each author of GNU Go get to enter his version with 
his changes?    If I make a little change to GNU GO and contribute
it to the project,  I get penalized for this?    But if I keep the
change to myself I get be an original work?

I don't think some of us are thinking clearly.

- Don












   David Fotland wrote:
   > Now let's suppose that instead of 10 copies of Gnugo, the tournament has 10
   > programs that are minor derivatives of gnugo with similar playing strength.
   > I think it's obvious that this is equally unfair.

   But what if the derivative is 4 stones stronger? How can you call
   that minor, even if only 1% of "time" was necessary? 

   Christoph
   _______________________________________________
   computer-go mailing list
   computer-go@xxxxxxxxxxxxxxxxx
   http://www.computer-go.org/mailman/listinfo/computer-go/

_______________________________________________
computer-go mailing list
computer-go@xxxxxxxxxxxxxxxxx
http://www.computer-go.org/mailman/listinfo/computer-go/