icsdrone(ng), polyglot + uci program

[flok]

Hi,

Anyone got an idea how to get the icsdroneng + polyglot combination to tell the (uci) program that it won or lost? Because it looks like it only sends a ucinewgame and that's it.

[Evert]

There is no standard way to tell an UCI program the result of a game.

[flok]

Why not introduce a method for doing so?

Code: Select all

uciresult 1-0
uciresult 1/2-1/2
uciresult 0-1
uciresult 1-1

[hgm]

If you want that kind of info, why not use a protocol that supports it, rather than resorting to non-standard methods that no other software would support?

Note that USI defines the command

gameover win|lose|draw

for this purpose, and that UCI2WB supports that. So if you want to make an adapter that translates the CECP 'result' command this way also in UCI, you would just have to remove the if-statement that limits it to USI mode fromUCI2WB.

[flok]

If you want that kind of info, why not use a protocol that supports it, rather than resorting to non-standard methods that no other software would support?

That no other software supports YET.

Note that USI defines the command

gameover win|lose|draw

for this purpose.

Yes but usi (assuming that that is the xboard protocol?) is far more work to correctly implement.

[hgm]

Yes but usi (assuming that that is the xboard protocol?) ...

XBoard protocol is CECP.

Note I was still editing my previous posting.

[flok]

Oh right, sorry.

What would be the path to get this into UCI?
I'll write an e-mail to the people that wrote uci.txt

[flok]

Ok I've implemented a patch for polyglot and sent that to the authors of polyglot. Let's see what happens.

[Evert]

Oh right, sorry.

What would be the path to get this into UCI?
I'll write an e-mail to the people that wrote uci.txt

Good luck.
In the past, suggestions for adding missing functionality to UCI have been met with great hostility.

[hgm]

UCI should be considered a dead protocol, that is set in stone.