Comments/Ratings for a Single Item

Very interesting.

Good initiative.

Maybe you could also add the Parlett's notation, sometimes it is more straight than Betza's. I find that both complete rather well. It is what is done in the WP page (on which I have contributed a lot recently).

https://en.wikipedia.org/wiki/Fairy_chess_piece

I have never heard of the Parlett notation before. I chose Betza notation for use in creating unique IDs for the pieces, because it is commonly used on this site. I do not believe Parlett notation is widely used here, and I need only one piece notation system for this purpose. However, if there is enough interest in Parlett notation from people who are willing to enter data for this project, I could add a column for it.

No problem, my suggestion can be ignored.

Dave Parlett is a famous author of books about games, also an historian. His books on card games are a reference. He is also the author of Hare & Tortoise a very original race game without any dice, only skill-based. He wrote an History of book games which contains a full chapter on chess variants and it is in this book that he proposed a notation which has some merits, at least for common fairy pieces. Quite a good book btw for anyone who would have interest in.

For most fairy pieces Betza notation is very compact, and would consists only of 2 or 3 capitals. Like BN and BR for Archbishop and Chancellor, HFD for the 'Half Duck', FC for the Omega Wizard. It gets a bit complex only for asymmetric pieces, where you would have to detail all the directions it moves in. And for divergent pieces, where you have to specify captures and non-captures separately (fmWfcF for a Shatranj Pawn, which is divergent as well as asymmetric).

With the XBetza extension it can get really complicated to specify moves that consists of multiple 'legs', i.e. visit intermediate squares on the way to their final destination to do 'something' there. (Which could be as innocent as just changing direction, or it could be testing whether the intermediate square is empty (lame leapers) or occupied (hoppers), or even capture what is there (locusts).) In that case you would have to specify the entire path to the destination by a sequence of (relative) direction specifications. Which will get longer as the paths visit more squares, and as their are more alternative paths. The Falcon of Falcon Chess is a notorious example of that: it has 16 move targets, each reachable through 3 different path, each path consisting of 3 steps. (Fortunately it is fully symmetric, which makes the 16 targets equivalent, but that still leaves 3 paths of 3 steps to be described.) Such pieces are really exceptional, though.

This is an interesting project. I'm not sure about using Betza notation for the primary keys though.

Although we can specify some rules, such as alphabetical order of discrete moves, I don't think that completely solves the problem of uniquely determing the notation for a piece and applying the rules isn't really "user friendly" (assuming people will be selecting these values - if it was only a database primary key, it could just be an integer.)

But a bigger problem than uniqueness is that, for some pieces, the Betza notation is so user-unfriendly that it is really only suitable for programming. For example, one notation for the Falcon is "afafsKafsafKafraflKaflafrK". Or consider the Advancer - the XBetza is a multi-leg move to move on top of the victim and then back off one space so it's pretty ugly too. I think "Falcon" and "Advancer" are far more user-friendly identifiers.

You have a table to allow specifying that pieces have different names in different games. I think we can have one standard name used for internal identification and, in most cases, it is pretty obvious what this internal name would be. In ChessV you can call any piece whatever you want in any game but each pre-defined piece has an "internal name" which is used for global identification. I generally had no problem deciding what internal name made sense. The only place I remember having some difficulty was the Griffin/Aanca pieces. For these I settled on calling them Ferz Then Rook and Wazir Then Bishop, which are admittedly non-standard, but at least very clear. Overall this approach has worked well. I have probably close to a hundred pieces defined - I'll dump a list of the names...

I think "Falcon" and "Advancer" are far more user-friendly identifiers.

Yes, it may be best to not use Betza notation for pieces like these.

I think we can have one standard name used for internal identification and, in most cases, it is pretty obvious what this internal name would be. In ChessV you can call any piece whatever you want in any game but each pre-defined piece has an "internal name" which is used for global identification.

Yes, the fairychess include file for Game Courier does the same thing.

One option is to use Betza notation where it is simple but to use a name where it gets complicated or can't be used. The other option is to use an internal name for every piece.

One option is to use Betza notation where it is simple but to use a name where it gets complicated or can't be used. The other option is to use an internal name for every piece.

Agreed.  So I guess the question is whether there is advantage to using Betza for some pieces.  Let's consider the Chancellor.  I guess the downside to using internal name is that some might consider it Marshal instead and might not recognize it as Chancellor (ok, maybe this piece wasn't the best example.)  OTOH, anyone understanding basic Betza notation will know this piece is either RN or NR but wouldn't know which and would need to check for both unless they knew and applied an additional rule that stipulates alphabetical order.

Overall, I'm doubtful that using Betza notation provides any real advantages, except that it could avoid some potential disagreement about what internal name should be used for a piece.  But I think we just state that this is internal bookkeeping data and isn't intended to say what a piece should be called.  And I don't think that end users of this data would necessarily even need to see the internal name, although people entering or maintaining the data would.

These are the internal names from ChessV:

Amazon, Archbishop, Bent Hero, Bent Shaman, Berolina Pawn, Bishop, Buffalo, Caliph, Camel, Camel General, Camelrider, Cannon, Centaur, Chained Padwar, Chained Warrior, Champion, Chancellor, Charging Knight, Charging Rook, Chu Shogi Lion, Cleric, Colonel, Copper General, Dabbabah, Dabbabahrider, Diamond Chess Pawn, Dragon Horse, Dragon King, Elephant, Elephant Ferz, Falcon, Ferz, Ferz Then Rook, Free Padwar, Free Warrior, Frog, General, Gold General, High Priestess, Joker, Jumping General, King, Knight, Knight Then Bishop, Kylin, Lightning Warmachine, Lion, Mao, Ministor, Narrow Knight, Nightrider, Oliphant, Pawn, Phoenix, Queen, Rhino, Rook, Scout, Short Rook, Side Mover, Side Mover General, Silver General, Sliding General, Squirrel, Squirrel General, Tower, Tribbabah, Tribbabahrider, Unicorn, Vao, Vertical Mover, Vertical Mover General, War Elephant, Wazir, Wazir Then Bishop, Wildebeest, Wizard, Zebra, Zebrarider

Looking at this list, there are some instances where the Betza notation could be easier to find.  For instance, the Camel-Bishop hybrid.  I use the name Caliph for this piece, but that would not be obvious to someone who hasn't seen that name used or doesn't remember.  But for occasional pieces like this, the internal name for it could be, for example, Bishop Camel.  I used this approach for the Elephant Ferz. In any event, we can just make a cheat sheet that lists out common names and betza notations when we run into occasional pieces like this where it isn't obvious.

Upon further consideration, I think I can populate a lot of this to get us started. I have ChessV configured for about 120 games and it knows what pieces are in each game and what they are called in that game. With a little cross-referencing, I should be able to export the data for those games in a usable format ...

I think You neam minister instead of ministor!

Indeed, now fixed. Thanks!

I have initially populated the PieceInfo table with data from the Item table for Piececlopedia pages and corresponding LinkDescriptions from IndexEntry, and I have begun to tweak this data for an initial listing of PieceIDs and descriptions.

Is that visible somewhere?

Is that visible somewhere?

At present, it is visible only to editors who have the password to the database. It will be visible to others in time.

I have written a script that will list the pieces in the PieceInfo table:

https://www.chessvariants.com/piececlopedia.dir/pieces.php

Each PieceID is linked to the Piececlopedia page with a /page/* URL, and each has a description, a category, and a variable number of attributes. The category is used for grouping them, and the category name appears in a heading instead of in the table. The description should provide a clear understanding of how a piece moves. Some attributes match categories, but others do not.

I am still working on selecting the terminology. I think Chimera is a good term for a piece that combines selected powers of different pieces without being a full compound piece. But I'm still unsettled on the term I should use for pieces that possess a subset of the powers of a particular piece. I described these with the Incomplete category and the Asymmetric attribute. However, the Asymmetric attribute also applies to Chimeras, and I don't really like the term Incomplete. The pages for the Barc and Crab describe them as restricted knights. I could use the Restricted attribute, but I was already using it to describe long-range pieces that have some restriction on their ability to move in the directions they are allowed to move. For example, the Edgehog moves like a Queen but with the restriction that its move goes from one edge to another. One possibility is to just extend the meaning of Restricted to allow it to also include an inability to move in a particular direction that its parent piece can move in. For example, a Lance could be described as a restricted Rook. I described the Hippogriff as Restricted, which moves as a Griffon with the restriction that it cannot go to the first three spaces in its path. Although this is not reducing the directions a piece can move, it is straightforwardly reducing the spaces it may move to. Thoughts?

How about 'dismembered' for completely lacking a move in a certain direction? For a slider that needs to travel a certain minimum numbers of steps, but still can be blocked there (the Tamerlane Picket is another example) you could use the qualification 'remote'. (If it would leap over the squares entirely it would be a 'ski-slider'. Which is a sort of degenerate case of a bent slider, where it doesn't change direction, but only stride.)

How about 'dismembered' for completely lacking a move in a certain direction?

No, that's a bit too graphic and emotionally loaded. I'm looking for something more neutral. Some terms I'm considering include fraction, fragment, moiety, segment, subcomponent, and subdivision. Perhaps subcomponent would be best. It is the most closely related to the word compound. A compound piece has its components, and a component could be split up into subcomponents.

I think "subset" would seem the most natural to me. I would also personally place the Silver General and company to be in this group rather than Chimera.

While thinking of the idea of Chimera pieces, and particularly of Derek Nalls' Zig-Zag pieces, I came up with one that could appropriately be called a Dragon. Since that name has already been popular, other possibilities are Winged Serpent or Chinese Dragon, for the idea behind the piece is the long serpentine Dragon depicted in China, not the western type that loves to hoard treasure. Wyvern is another possible name. There are four static variations on this piece, and a rotatable version that could switch between any of these four. There are also eight static subvariations of this piece and a rotatable version that could switch between the eight. The basic idea is to have a linerider along a single orthogonal or diagonal axis, which can move to either side as a Knight. The available Knight moves would be those that do not end up adjacent to the piece's axis. For an orthogonal axis, these would be ones that can be reached with an orthogonal move perpendicular to the axis, followed by an outward diagonal move. For a diagonal axis, these would be ones that can be reached with a diagonal move perpendicular to the axis, followed by an outward orthogonal move. So, whatever the orientation of the piece, it would have up to four possible Knight moves. The axis along which a Rook or Bishop move is possible would represent the serpentine body of the Dragon, and the Knight moves would represent its wings. For simple record keeping, I could call the four static versions the Dash Dragon, the Pipe Dragon, the Slash Dragon, and the Backslash Dragon. These names are based on punctuation marks that correspond with the orientation of the piece, but they may not be poetic enough for an actual game. I'm thinking of two different versions of the rotatable version. In one version, it may rotate to any other orientation after moving, or in place of moving. In a slightly weaker version, it could rotate 45 degrees either way after moving, or rotate to any other orientation in place of moving. Rotation should not be allowed before moving, for that would make the piece equivalent to an Amazon. If we conceive of the Dragon as having a head at one end and a tail at the other, we could have eight static types with only one radial direction of movement. A rotatable version would be able to switch between orientations after moving, or in place of a move. Since allowing this piece to rotate before moving would make it an Amazon, this would not be permitted. Stronger variations, in which the piece has Nightrider moves instead of Knight moves are also possible.

I think "subset" would seem the most natural to me.

After sleeping on it, I woke up feeling that subset would work.

I would also personally place the Silver General and company to be in this group rather than Chimera.

With respect to the pieces in both Shogi and Chess, these are both subsets of the King. However, when I adapted these pieces for Hex Shogi, I conceived of the Gold General as moving like a Wazir or forward as a Ferz and of the Silver General as moving like a Ferz or forward as a Wazir. These are accurate descriptions, which may both be regarded as Chimeras of the Wazir and Ferz. But it is also true that the Hex Shogi versions are still subsets of the Hex Shogi King.

If we're to put these into a Subset category, then this would work best if the Man were put into the Simple Leaper category rather than the Leaper Compound category. Is that something you would be comfortable with?

If we're to put these into a Subset category, then this would work best if the Man were put into the Simple Leaper category rather than the Leaper Compound category. Is that something you would be comfortable with?

Yes, that's where I would put it.

The Man/King is not a simple leaper. That term is reserved for pieces where all moves are rotated or mirrorred versions of each other. So the King is definitely a compound.

Compounds can have subsets to, though. Which then can be compounds of subsets of simple leapers. So I don't think that we can require a piece must be one or the other. Subset merely is an extra qualification that can be applied to simple leapers and compound leapers alike. (Or to sliders, for that matter.)

So there doesn't seem anything wrong with saying Gold is a subset of the King. That is more true than just saying it is a compound, as it is really a compound of subsets. That the King is a compound should be considered common knowledge. King and Queen have special status in this respect, because they are orthocox Chess pieces, and people can be expected to know them. Betza notation expresses their moves as a single capital, K or Q, officially 'shorthand' for WF and RB.

The large Shogi variants have many pieces, and they are virtually all subsets of the Queen. (As oblique moves are virtually non-existent in Shogi.) Some are even subsets of the Rook or the King.

What I dislike about the term 'subset' is that it has a well-defined meaning, which would consider the R4 a subset of the Rook. It seems that we now want to deviate from this meaning, by not allowing it to describe range restrictions, but only complete absence of a move in a certain direction. I think that such a distortion of the meaning of well-known terms is inadvisable.

I guess I am persuaded that a Man is a compound, since a Queen is, but that a Gold General can still be considered a Man/King subset.

I'm not persuaded, however, that an R4 would be a R subset. It depends what we are referring to as the items in the set. I consider it a set of move capabilities, not a set of potential destination squares. And I consider a range-limited slide a different move capability than an unlimited slide. I certainly don't consider a dabbabahrider a subset of a rook.