CRoberson wrote: Yes, moding the format to a strict standard screws all the UCI folks using an interface that doesn't comply.
I am pretty sure most interfaces would comply with a standard if you set one well-enough in advance. And they can always use Polyglot, can't they?
Guests can watch tournament games as long as the games are unrated which is why our games are unrated.
I tested that myself during the last tournament.
What do you mean by "watch"? Observe or follow?
I want to collect everything from the engines not just the last kibitz for a given move.
Awful! I sure hope you will not put all that in the PGNs. Or at least make PGNs avalable that have simple score/depth embedded as well.
And the original question still stands: If you allow (and collect) multiple kibitzes per move, how are you going to figure out which kibitz message belonged to each move?
hgm wrote:
What do you mean by "watch"? Observe or follow?
Awful! I sure hope you will not put all that in the PGNs. Or at least make PGNs avalable that have simple score/depth embedded as well.
And the original question still stands: If you allow (and collect) multiple kibitzes per move, how are you going to figure out which kibitz message belonged to each move?
I meant observe since one can observe multiple games per session.
I think you are missing the point. I think this method may be able to replace the need for real-time log file submission. Not considering PGNs at this time.
hgm wrote:
What do you mean by "watch"? Observe or follow?
Awful! I sure hope you will not put all that in the PGNs. Or at least make PGNs avalable that have simple score/depth embedded as well.
And the original question still stands: If you allow (and collect) multiple kibitzes per move, how are you going to figure out which kibitz message belonged to each move?
I meant observe since one can observe multiple games per session.
I think you are missing the point. I think this method may be able to replace the need for real-time log file submission. Not considering PGNs at this time.
The problem is that observe is time-sensitive. If you observe before the first move is played, you can capture everything. But if you are even one move late, the PGN will be broken. Follow would be much nicer. And one could _still_ make this work with follow, if you can set up N ICC accounts. A single application can open N telnet connections, and have each connection follow one participant, which would capture things from move 1 correctly.
hgm wrote:Does that mean you are going to insist on a format then that makes it impossible to get good PGN, just because PGN does not interest you?
If this code has to be written, it can be made to work. The "observer" software will know which move is played as the game progresses. It can remember the last PV kibitzed that started with that move (kibitzed by that opponent of course, since some choose to ponder in oddball ways and appear to be searching for their opponent rather than themselves).
