GTP in tournaments (was Re: [computer-go] GTP and Tourney in SmartGo1.4)

[Marco Scheurer <marco@xxxxxxxx>]

On Mar 12, 2004, at 12:03 AM, Nick Wedd wrote:
> Insisting on GTP will exclude most competitors, as they don't support 
> it.
This is certainly an issue...

> I have heard that it is hard to implement, with two line-feeds having 
> a different meaning from one line-feed (my personal opinion is that 
> significant whitespace is a really stupid idea.)
>
> About five years ago, when many programs claimed to support GMP, the 
> chance of two such programs actually being able to play an entire game 
> by GMP was rather small.
... However, GTP is IMHO the best protocol available today to link go 
programs, and it's time to get rid of GMP. It's also not hard to 
implement.

A benefit of implementing GTP for the authors of the Windows commercial 
softwares could be to sell their engine to the small but growing Linux 
and Mac (unless they use assembly of course) crowd with little or no 
effort: they would work with a number of free GTP UIs available on 
these systems.

Markus Enzeberg has developed a set of tools that would help anyone get 
up to speed with the GTP and test it's engine 
(http://www.markus-enzenberger.de/compgo/gogui.html ). I think these 
tools also provide all the necessary machinary to hook programs running 
on different machines with a referee.

> If two of your entrants fail to get their game started, or play part 
> of a game and then go to sleep, do you propose investigating which one 
> is guilty of not doing the GTP stuff right, and giving the win to the 
> other?
Yes, that could be a difficulty. However, using a 3 machine setup with 
one acting as a referee (as I understand is possible with the above 
tools), it would be easy to determine which of the two players is not 
responding.

If not, the harsh and bureaucratic thing to do would be to ask each 
program to pass a regression test before entering the tournament. The 
unformal, friendly thing would be to allow other communication for 
instance a human transcribing the engine moves on a GTP capable UI. Or 
maybe even dump the requirement.

Last year the GIFU challenge (http://www.gifuit.com/igoe2003/) asked 
entering programs to implement a new protocol, the SGMP (to the dismay 
of the GTP fans). They included a way for entrants to test their 
program before hand. Did it work? All the top programs were playing at 
this tournament, so I suppose they complied and implemented the 
required protocol. At least I haven't seen any complaints about this 
requirement.

It may seem gratuitous to require the programmers to implement yet 
another protocol, but I believe that everybody would benefit if that 
one took hold.

marco


Marco Scheurer
Sen:te, Lausanne, Switzerland   http://www.sente.ch

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