🕸Fergus Duniho wrote on Thu, Feb 4, 2021 02:23 PM UTC:
I have started work on creating a piece database. At this point, I have created three tables, and I will be working on a script for updating them. The three tables are called PieceInfo, PieceImages, and PieceNames.
Each one includes a column called PieceID. For simple consistency, I expect it makes sense to use Betza codes, perhaps with some added standardization when two different codes could otherwise be used for the same piece, such as alphabetizing parts that could be written in an arbitrary order. Since I don't work with Betza codes much, others may have more insight on this. Presently, this column can be up to 32 characters in length. That's just a guess, since I don't know how long Betza codes typically get.
PieceInfo will contain a description of how a piece moves and the ItemID for the Piececlopedia page on that piece. This will be a one-to-one table.
PieceImages will be a many-to-many, or maybe many-to-one, table in which PieceIDs are associated with URLs to images stored in /graphics.dir/. I should probably include separate columns for White and Black images.
PieceNames will be a many-to-many table in which piece names are associated with PieceIDs. This will be indexed by a combination of PieceID and ItemID, where the ItemID will be for a page describing a game using that piece. The Name column will store the name used for the piece on that page.
The purpose behind this database is to build a better and more comprehensive Piececlopedia by collecting extensive data on piece usage and naming. Also, the PieceImages table may prove helpful for Game Courier or the interactive diagrams.
Since I am not versed in Betza code, and this is a big project, I am thinking of writing code for crowd-sourcing the work and for mining the interactive diagrams for data, since it already includes much of the requisite date in standardized forms.
I have started work on creating a piece database. At this point, I have created three tables, and I will be working on a script for updating them. The three tables are called PieceInfo, PieceImages, and PieceNames.
Each one includes a column called PieceID. For simple consistency, I expect it makes sense to use Betza codes, perhaps with some added standardization when two different codes could otherwise be used for the same piece, such as alphabetizing parts that could be written in an arbitrary order. Since I don't work with Betza codes much, others may have more insight on this. Presently, this column can be up to 32 characters in length. That's just a guess, since I don't know how long Betza codes typically get.
PieceInfo will contain a description of how a piece moves and the ItemID for the Piececlopedia page on that piece. This will be a one-to-one table.
PieceImages will be a many-to-many, or maybe many-to-one, table in which PieceIDs are associated with URLs to images stored in /graphics.dir/. I should probably include separate columns for White and Black images.
PieceNames will be a many-to-many table in which piece names are associated with PieceIDs. This will be indexed by a combination of PieceID and ItemID, where the ItemID will be for a page describing a game using that piece. The Name column will store the name used for the piece on that page.
The purpose behind this database is to build a better and more comprehensive Piececlopedia by collecting extensive data on piece usage and naming. Also, the PieceImages table may prove helpful for Game Courier or the interactive diagrams.
Since I am not versed in Betza code, and this is a big project, I am thinking of writing code for crowd-sourcing the work and for mining the interactive diagrams for data, since it already includes much of the requisite date in standardized forms.