Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Thanks, Bas !

Author: José Carlos

Date: 01:41:47 08/22/02

Go up one level in this thread


On August 21, 2002 at 19:23:33, Bas Hamstra wrote:

>On August 21, 2002 at 18:05:14, Russell Reagan wrote:
>
>>On August 21, 2002 at 17:33:05, Bas Hamstra wrote:
>>
>>>You are not the only one of course, I remember having problems too, and NEVER
>>>have been 100% satisfied EVER about my WB code. I got it to work only after I
>>>got stuck a couple of times first, and finally realized it was a tough job and
>>>it simply can't be done right in a couple of hours. Got a command. Wait, AM I
>>>pondering? If yes, do I have to stop the search? Oops, it pondered to mate, it
>>>is not searching. Different case. But uhm, do I have to retract that pondermove?
>>>Depends on the command. Mindboggling.
>>
>>Yes, not everyone is intelligent enough to implement this kind of super complex
>>protocol where you have to keep track of more than 10 variables. Absolutely
>>mindboggling...
>>Maybe the new slogan for UCI should be, "The protocol for the non-programmer".
>
>>>WB is great of course and a *immense* success. Best free app ever and I hope
>>>they keep maitaning the source. But UCI shows that some things can and should be
>>>done really more simple and elegant.
>>
>>Do you not understand the purpose of such a protocol? The purpose is to provide
>>a standard, graphical way for a text based engine to communicate with a user, >another engine, or an ICS. The protocol should not be intrusive upon the
>>engine, at all.
>
>That's about the 60th time you use that word. Tell me (in you own words if
>possible, not Bob's) what you exactly mean by it? What is your problem? An
>engine is just a searcher.

  No it isn't. An engine does many more things when playing a game, than just
search.
  An engine is just a searcher in the case you use it to analyze under CA or CB,
but that's not the only use of an engine.

>You tell it: search! And there it goes. However,
>without any intrusion, it won't search.

  This is wrong. You don't need to tell the engine what to do, far less how to
do it. You just have to give _information_ to the engine, and the engine will
decide what to do. What the engine demands from the GUI is "what's happening
outside my range of view? what does the user want from me? what is the situation
of the game?..."
  Again, the exception is when analyzing under CA, CB, etc.

>It just sits there. And I don't know
>about you, but if *I* were an engine, I would LOVE some intrusion.

  Try to imagine this applied to human players. "Now you play this move out of
book" "Now you ponder on Bc4" "Search!". Sounds ridiculous.

>Because
>without any intrusion I couldn't do what I was best at: searching. But alas, no
>intrusion without protocol. You could try to intrude without protocol, but
>chances are the engine wouldn't understand the meaning of the intrusion. So the
>whole discussion is more or less what is the best way to intrude the engine,

  "Inform" the engine, not intrude it.

>you understand? If you have any questions, don't hesistate to ask!
>
>Best regards,
>Bas.

  I have one question, though I don't think you know the answer: why wasn't UCI
protocol discussed here _before_ considering it definitive?

  Regards,
  José C.



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.