BertTuyt wrote:
Also the 7P Database is free for all.
The problem is the distribution, as it is 30 GByte.
There are some options for this, like uploading this to a cloud based service (and if we wait a little than also 30 GByte is for free).
If I remember well the 7P Database from Michel is much smaller, I don't know if he used a radically different compression technique (maybe he is willing to share this).
One simple way of reducing the size, is to exclude the positions where site to move and other side to move have a capture.
Nowadays only the side to move capture situations are excluded.
This will make the compression more efficient.
Dragon's current databases are compressed in the way you describe. It saves considerably (factor 2 or so) on space requirements.
Dragon's databases can be downloaded from
http://mdgsoft.home.xs4all.nl/draughts/endgame.htm
A lot of databases are omitted because they are pretty much useless; 1 man+ 3 kings vs 1 man+3 kings is either evident with a 2 ply search, or it is a draw.
BertTuyt wrote:
Another approach but that takes time, and I'm not sure that we will get to this situation , is also to standardize the EndGame Database format.
...
But again im not sure to what extend programmers can agree upon this approach.
Creating a DB (in comparison with a game database, see another discussion in this forum) is mainly computer extensive, for which one need to have the right resources.
Creating 1 standard would have it's merrits, but it would also stiffle innovation. The standard would have to encompass things like distance-to-win, partial databases, or damy's draw+ databases, and different compression schemes.
How about a database of 5 man vs 5 man, based on partial results, with all man between fields 10 and 40? Possibilities for innovation are endless.
Michel
