bug-gnubg
[Top][All Lists]
Advanced

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

Re: Re: [Bug-gnubg] window size


From: Massimiliano . Maini
Subject: Re: Re: [Bug-gnubg] window size
Date: Tue, 2 May 2006 16:08:58 +0200

>On Tue, May 02, 2006 at 02:48:34PM +0200, Øystein O. Johansen wrote:
>>
>> The evaluation window is really old and untouch code. It uses the old
>> widget (from the pre gtk2.x era) called GtkText. (GTK2.0 and later
>> should use GtkTextView instead or ... ) It has not been touched, since
>> it's not a window that's used much. We've also several times discussed
>> to remove the window. Some of the information in this windows is even
>> outdated and not correct. The 'value' and 'dE/dI' numbers are
>> incorrect and gives no meaning what so ever, and the other values can
>> be better found with 'hint'?!?
>
>Then let's remove it. GnuBG uses far too much deprecated gtk code and if
>the values given in the window are wrong as well as useless, I see no
>reason to keep the evaluate function.

I may be a dreamer but I've been wondering for a while if it was possible
to extract useful information from the eval window data (not knowing it's
crappy at the moment).

I mean something like : move #1 is better than move #2 because of the race,
or because of the gammon chances, or because of it keeps the contact, etc.

This corresponds to pull out some qualitative information from the neural
net outputs ... not really easy I suppose, but this would be the ultimate
learning tool ... a tool that explains you why you suck (instead of just
telling you that you suck) :))


MaX.






reply via email to

[Prev in Thread] Current Thread [Next in Thread]