Scidb binary file format

[dkl]

Is there any technical specification of the Scid database format? Like a real technical specification, not just some informal description like http://scidvspc.sourceforge.net/doc/Formats.htm .

Or do I have to do the tedious task on crawling and understanding the source code? (In some sense, Yarin's reverse-engineered docu http://www.talkchess.com/forum/viewtopic.php?t=29468 for CBH is easier to implement against than trying to get through some very specific and not always very commented source-code...)

[gcramer]

Is there any technical specification of the Scid database format?

No, the database format is still in development, a technical specification will not be written before it has reached a final stage.

In some sense, Yarin's reverse-engineered docu http://www.talkchess.com/forum/viewtopic.php?t=29468 for CBH is easier to implement against than trying to get through some very specific and not always very commented source-code...

Probably Yarin's docu is better to understand, but Scidb's implementation has solved some subtleties (not all) which are not documented in this docu. So it is recommended to analyze Scidb's source code, but for a start Yarin's docu is better; in a later stage, when you've understood the CBH format, it will be much easier to understand Scidb's source code. Keep in mind that Scidb's source code is also still in development, and contains some experimental code, reading CBH databases is a trial and error task.

[dkl]

Is there any technical specification of the Scid database format?

No, the database format is still in development, a technical specification will not be written before it has reached a final stage.

I was not thinking .scidb, but rather .si3 and .si4 which seem pretty stable to me.

In some sense, Yarin's reverse-engineered docu http://www.talkchess.com/forum/viewtopic.php?t=29468 for CBH is easier to implement against than trying to get through some very specific and not always very commented source-code...

Probably Yarin's docu is better to understand, but Scidb's implementation has solved some subtleties (not all) which are not documented in this docu. So it is recommended to analyze Scidb's source code, but for a start Yarin's docu is better; in a later stage, when you've understood the CBH format, it will be much easier to understand Scidb's source code. Keep in mind that Scidb's source code is also still in development, and contains some experimental code, reading CBH databases is a trial and error task.

I was bringing this just up as an example of what kind of document I mean by "specification". Scidb does an excellent job - afaik the only free tool capable at all - at reading respl. converting CBH, I do not think there is any point in "cloning" such functionality, when scidb works...

[gcramer]

Ok, a misunderstanding, because the subject is "Scidb binary file format". I don't know a specification of this format, but the source code is relatively easy to understand.

[dkl]

you're right, that title is kinda misleading and not very well chosen, but I couldn't edit it anymore...