Author: Lex Loep
Date: 10:39:52 02/02/02
Go up one level in this thread
On February 02, 2002 at 09:15:03, Odd Gunnar Malin wrote: >On February 02, 2002 at 03:13:17, Lex Loep wrote: > >>On February 01, 2002 at 14:01:24, Odd Gunnar Malin wrote: >> >>>On February 01, 2002 at 13:06:51, Ed Schröder wrote: >>> >>>>On February 01, 2002 at 07:54:51, J.Dufek wrote: >>>> >>>>>On February 01, 2002 at 05:14:15, Ed Schröder wrote: >>>>> >>>>>>On February 01, 2002 at 03:37:58, Dann Corbit wrote: >>>>>> >>>>>>>On February 01, 2002 at 02:57:21, Ed Schröder wrote: >>>>>>> >>>>>>>>On February 01, 2002 at 02:00:02, Jose Menendez wrote: >>>>>>>> >>>>>>>>>/ >>>>>>>> >>>>>>>>http://www.rebel.nl/ubb/Forum1/HTML/000379.html >>>>>>> >>>>>>>How soon is soon on the Winboard Rebel? >>>>>>> >>>>>>>As you know, I have been panting for this for ages. >>>>>>> >>>>>>>Generally speaking -- 1 month? 1 quarter? >>>>>> >>>>>> >>>>>>Probably in between. >>>>>> >>>>>>Ed >>>> >>>> >>>>>Hi Rebel's group! :-) >>>>>Very good message. >>>>>One question for Ed: you wrote about enhanced winboard protocol. But now coming >>>>>Arena :-) - will be a possible run Century with all features (CAT, book ...) >>>>>under this GUI (or how about implementing UCI ...:-) ) >>>>>Thank you very much for your answer >>>>>Best regards >>>>>J. >>>> >>>>I heard many good things about Arena in the between time, I haven't had the time >>>>to have a look at it. Where can one download it? >>>> >>>>About Rebel: the interface will support UCI engines and an extended Winboard >>>>protocol which allows every WB programmer to add some basic missing WB features >>>>such as the display of opening book moves, the current line "Move Nf6 (7/44) >>>>depth=14", flexible hash table size etc. >>>> >>> >>>You are talking of the Interface could show this info and not that you have >>>extended the protocol to have this info? >>>I ask because both bookmove (bk) and current thinking move (stat01) is already >>>in the protocol and suported by many engines. >>>The output from stat01 line is prety standard but the bk output is of free >>>format, have you made an output standard that the interface need. >> > >There was propose by Dieter Bürßner to have a GUI command in the protocol so the >GUI could ask the engine if it have implemented some special command for this >GUI. I think some GUI programmer should take this up in the right fora. > >>According to the WBII specs stat01 is only defined for analyse mode and > >I tested some engines with this command in normal search an most of them would >give the stat01 message, a few was ignoring this under search that would maybe >be the correct way strictly to the protocol. A small number was actually taking >this as a 'move now' (?) command but this engines would also move if they get a >draw offer under search so to please them you must have a toggle switch in the >engine setupbox that could stop all communication except for 'move now' and >maybe 'quit'. >We are not talking of the maine use of the program here, playing against an >engine, because I don't think many would like to see any search information >then. >So this let us to a few engin-engine runners. >Another main use of a program is analyze but here the stat01 line is allowed. > That is exactly the problem, the command is not well defined and the implementation is erratic. I still wanted this information so I defined it such that it would not break the standard. >>having a bk command with an undefined output is pretty useless. > >You could show a box with this information but I agree this isn't much in a >commercial GUI. Here you must be allowed to follow the tree by clicking on a >move. >I'm not sure what you would do with this information, if you should be able to >alter any information in the book there must be a book-mode commandset too. >Have you added a 'addbookmove <move> <score>' command or something. > >>Further more there is no defined way to set various engine parameters like >>hash size, tablebase path, books, personalyties, styles etc. > >I agree this make the engine hard to use by 'none computer experts'. The >protocol need some changes in this area. > >>To solve al these I added just two commands: >> >>setvar - sets a variable >>getvar - to retreive a variable. >> >>The list of variables is currently in a ini file which belongs >>to the engine, the GUI uses the list to build a dialog and to >>send the proper commands to the engine. > >I first thought we was back to stage one here with the user fidling with the ini >file but you of course talk about a 'one time' ini file that the author or >another who knows the engine could make and would follow the engine. > Exactly, it would be ideal if the engine authors supplied such file with the engine. I know some go to lots of trouble to adapt to the crappy chessbase winboard adapter, it should be trivial to write a simple ini file. >It is ok with changes but I would love if this also was discussed in the proper >fora first. >But I think it isn't so wrong if you don't broke any compatibility with the >standard. > Nothing is broke. Lex >Odd Gunnar
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.