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

Re: [computer-go] I know we disagree,but I choose to do nothing about it.

In message <200507230118.10804.drd@xxxxxxxxxxxxxxxxx>, drd@xxxxxxxxxxxxxxxxx writes



I thought, I still think, that it is you who advocates mandating it. You
don't propose mandating it by requiring programs to respond in some way
to the "kgs-genmove_cleanup" command;  you advocate effectively
mandating it by changing the rules of the tournaments so that programs
that don't respond appropriately to it will lose most of their games as
a consequence.

And I am not happy about this.  Or maybe I still don't understand.



Nick,

Yes, no matter how you slice it,  I advocate mandating the protocol, even if
it's implied by the server and not required by the go program.




Don,



Now we are getting somewhere.  Not with the main argument, but with my 
understanding of what the argument is about.




But I'm not married to either viewpoint.   If you think it's cleaner to
REQUIRE the commands,  then I will concede to your judgement.   There are
some good arguments for both points of view.




I do not think it is cleaner to require the commands.  I have never 
thought this.  I used to think that you thought it;  now I see that I 
was wrong in this.




In this case, you can't say that a program was unfairly punished because it
failed to respond to the commands.     Someone might claim unfairness because
it responded stupidly, but that is like claiming unfairness because my
program played a bad move.




I wouldn't say it was unfair, if the rules were the same for all.  But 
that is not a route which I wish to follow, at least for a while, for 
reasons I have already explained.



There are three paths we can follow:



(I.)  Continue as at present.  Bots will be able to support the game-end 
protocol if they like;  but if after successive passes they fail to 
agree on the score, and at least one fails to support the protocol 
correctly, the game is scored by human intervention.



(II.)  Encourage support for the game-end protocol.  No more manual 
intervention.  Bots which fail to support the protocol will lose many 
games as a consequence.



(III.)  Require support for the protocol.



I advocate (I.), with maybe a transfer to (II.) in the future.

I now believe that Don advocates (II.)

So far as I am aware, no-one has ever advocated (III.)



Nick

--

Nick Wedd    nick@xxxxxxxxxxxxxxxxx

_______________________________________________

computer-go mailing list

computer-go@xxxxxxxxxxxxxxxxx

http://www.computer-go.org/mailman/listinfo/computer-go/