| Author |
Message |
Mark Thompson (Markt)
| | Posted on Monday, January 06, 2003 - 6:50 am: | |
I think it would be great if the ZSG files could include "comments" that would amount to annotations for the game. There would need to be a window that could be shown while playing where one could enter a comment, and it might include a check-box for letting the other player see the comment before the game is finished. Another window would show the comment when the game is reviewed. |
David J Bush (Twixter)
| | Posted on Monday, January 06, 2003 - 9:27 am: | |
Would this comment window be separate from the Zillions window? All the other windows (move list, chat, etc.) have been glommed onto the main window. There's not much room there for added text. A separate window might be more confusing than helpful for many users. Annotations are usually made after a game, not during. Maybe the chat window could be used after the game for entering text relevant to specific moves. This text could then be saved with the game. Of course, better still would be the ability to store and examine an entire move tree of variations, not just the single line of the game. This would make puzzle solutions much clearer, for example. Many puzzles, such as chess puzzles, have lots of interesting variations. That's what makes them worthwhile. Just one winning line may not be enough to show the user why that way works and his/her idea doesn't. As long as we're turning Zillions into a game database manager, we should also include the ability to evaluate a "leaf node" as winning, advantage, draw, disadvantage, losing, or unclear. Of course Zillions should then be able to automatically label all the other nodes based on all the leaf nodes that node generates. |
Matthew Burke (Morat)
| | Posted on Monday, January 06, 2003 - 11:10 am: | |
Along these lines, it would be nice to be able to have the saved games written out in a more standard notation.
As for game annotation capability, a better solution might be to have a separate tool to do that. As long as we're wishing....how about having the Zrf parser and move generator sold as a separate product? |
Alfred Pfeiffer (Apf)
| | Posted on Monday, January 06, 2003 - 1:15 pm: | |
Eventually Zillions should use the syntax of the zrf-language also in zsg-files and ignore all characters until end of line when a ';' occured. This would allow to add any comments to the zsg with an usual ASCII editor.
With a such feature the next wish for Zillions of Games would be a possibility (command) to put texts (as comment) into the moves list. This then also could help when pogramming new zrfs. |
David J Bush (Twixter)
| | Posted on Monday, January 06, 2003 - 1:19 pm: | |
A more standard notation seems like a good idea, but there are many potential problems in its implementation. How would the user define this new notation? Seems like you would need to add lots of new features to the scripting language, just to be able to do that. For example, in chess the castling move is usually indicated by O-O or O-O-O. How could you tell Zillions to substitute "O-O" for "King e1-g1 Rook h1-f1" ? for a worse case, in Twixt the standard notation for a move is usually to just give the coordinates where the peg is placed, followed by an asterisk for each link that is automatically added as a result of that move. How could you tell Zillions to use "J7**" instead of "White Peg J7 White Peg jk06 White Peg JKZ6 White Peg JKZ5 White Peg jk08 White Peg JKZ7 White Peg JKZ8 = L1b on JKZ6 = Link2 on jk08 = Link1 on jk06 = L2t on JKZ7 = L1t on JKZ5 = L2b on JKZ8" ? That would be a lot of work for dubious reward. How many people would really want a text record of their game, As long as they have software that can examine the move tree? Zillions already saves games in a quasi-readable text format.
Maybe a Zillions database manager should be a separate product, especially if it does the text translation you suggest. But my hope is that a decent game tree analysis capability would be added to Zillions itself, for the sake of the puzzles if nothing else. Here are some more features I thought of:
* Define a specific line as the main (or game) line
* Buttons (or shortcut keys) added: Back to branch, Back to main line, forward to branch, forward to end, forward to end of main line, etc.
Perhaps the SGF format could be used? Here's a link: http://www.red-bean.com/sgf/
As far as I can tell, the only reserved symbols in SGF are ;()[] Any game that uses one of these symbols in its move syntax might conflict with SGF. But most games don't. SGF would allow other software (such as ZSG syntax parsers and translators) to interface more easily. |
Matthew Burke (Morat)
| | Posted on Monday, January 06, 2003 - 7:09 pm: | |
I don't know that it would be too hard. Add an optional string parameter to the "add" command that would be the notation to write out. With a reasonably thought out set of "special" characters (i.e. "%f" for the from square, "%t" for the to square, "%p" for the piece name, etc.) you should be able to fairly easily manage a wide range of notations.
So in the code for castling, you would have
(add "O-O") or (add "O-O-O")
and normal moves in chess could have
(add "%f-%t")
whereas pawn moves could be
(add "%p-%t")
etc. (bear with me, I don't actually play much chess so I'm a little rusty on the notation).
Funny you should mention SGF. My working with that is what prompted me to suggest having a separate program to do game analysis. There's a lot of headaches associated with deciding how to display variations, etc. and since ZoG is good at what it does, I don't know that it's worth adding all the overhead. Particularly since there's a long list of features related to preparing illustrations of game positions, etc. that I would really like to have. |
David J Bush (Twixter)
| | Posted on Monday, January 06, 2003 - 7:36 pm: | |
| I don't know that it would be too hard. Add an optional string parameter to the "add" command
| that would be the notation to write out. With a reasonably thought out set of "special"
| characters (i.e. "%f" for the from square, "%t" for the to square, "%p" for the piece name, etc.)
| you should be able to fairly easily manage a wide range of notations.
Sorry, I should have read your post more closely. There are cases where this would not work; for example, the code for a move that may or may not generate cascade moves, would have just a single add for all cases. Also, captures or regular moves could be represented by the same add. If a Twixt peg drop is going to have an asterisk for each automatic link it generates, how would a single string represent that? You might be able to produce something closer to human-readable with your approach, but there may be moves where the resulting notation would be ambiguous.
Certainly, it would be a lot of work to add move tree handling capability to Zillions. But it wouldn't be necessary to change the move notation per se, in order to do it. SGF could probably be added to the existing zillions syntax. |
Matthew Burke (Morat)
| | Posted on Monday, January 06, 2003 - 9:10 pm: | |
Hmmm....your post changed between my reading it in my office and my coming home and sitting down to respond. I didn't realize you could do that.
Good point about the cascades, etc. Even after a couple of years off and on doing zrfs (more off than on), I still don't really have my mind wrapped around it.
I'm not clear on what you mean by "SGF could probably be added to the existing zillions syntax." |
David J Bush (Twixter)
| | Posted on Monday, January 06, 2003 - 10:06 pm: | |
In the Utilities folder on the left, if you click on Edit Profile, you can edit your recent posts. Good thing, too; I made a total ass of myself. Not that that's anything new.
Actually, the example I gave of possible ambiguity was not a good one; for example, on the Twixt play-by-email server, the notation used dispenses with extra symbols for any automatic link. It is just assumed that any links that can be made, are made. Such notation is not ambiguous, just a little cryptic perhaps. In general, cascades could probably be determined from the move that generates them. So, your idea might do a decent job in most cases.
All I meant about SGF (about which I know very little) was, as long as the move syntax for some Zillions game does not use any of the symbols ;()[] then I suppose it could be turned into SGF notation. Some sort of addition like SGF would be needed, if Zillions were to handle move trees. |
Mark Thompson (Markt)
| | Posted on Monday, January 06, 2003 - 11:19 pm: | |
Re: annotations, some games have pretty tall board images that don't leave much vertical room for displaying notes. Hence I think of a separate pane, like the moves list, which could be displayed or hidden at will. I agree the notes should be edit-able after the game is over, ideally (though this could be done with any text editor), but personally I think it would be good to be able to scribble down my thoughts during the game, which would make it easier to construct more formal annotations afterward. You wouldn't have to guess what the players were thinking if they had recorded their own notes.
The idea of having separate lines of play is terrific, but would involve more new programming than what I was thinking of. On the other hand if you could record the lines of play you were anticipating with your move, that would greatly enhance the clarity of the notes.
Re: getting the ZSG's into more standard notation, I'd suggest writing separate programs for each game to translate the ZSG as desired. A special-purpose program can better deal with the idiosyncrasies of particular games, and most Zillions users probably aren't too concerned with translating their ZSG's anyway.
Re: Matthew's suggestion of having the ZRF parser and move generator available separately ... that would indeed be paradise, and would make a notation-generator (as for Chess) easier to write (since a move-generator would be needed to know whether it was necessary to say Nd4-c6 or could get away with Nc6).
I'd also love to get a parser/move generator in order to work on separate applications. I'd be interested in trying to create software that would generate "White to move and mate in 2" problems in Chess, and with a ZRF parser/move generator it might be possible to make such a program work on other chess variants also. (Assuming it could be made to work even on Chess.) |
Fergus Duniho (Fduniho)
| | Posted on Thursday, January 09, 2003 - 7:01 pm: | |
Speaking of annotating games, I just updated my web-based Chess Variants PBM server to handle them. Annotations can be added as C++ style comments, and it can also handle branching variations of play.
It can handle games, problems, tutorials, and even whole opening books. For more details and to see it in action, go to
http://play.chessvariants.com/pbm/help.html
Read the last two sections of the help file and look at the examples. |
|