Last week Gregory and his wife visited me here in Cologne. It's about 3 years ago that we first met in person in Seattle. Another 3 years that I contacted him for the first time via email, because I was looking for some chess control I could easily use to make my idea of a chess software which uses a flash-card concept coming true.
He is the mastermind behind the chess control, the parsing of PGN files and engine supports. So, quite a few hot topics for future versions.
I invited 8 dudes from 8 different regions of Germany to give him a warm welcome:
That was before he friendly reminded me that he doesn't drink any beer at all! Shoooot!
Anyway, poor Greg thought he will just stay a few days with us before he continues his journey with his wife through South Germany and Austria. Little he did know. I lured him into my CPT studio and quietly locked the door from outside. Telling him he has first to finish the UCI support stuff before his journey will go on. Lucky him he had a little SD chip with him and easily plugged in the new stuff. He is almost done and promised to finish right after he returns to Seattle.
Then he explained to me how to develop 3-times faster:
I quickly launched CPT and showed him that the UI of the chessboard doesn't match the overall UI quality of CPT 4 anymore.
He agreed. While it is not very likely that it makes it already into CPT 4 a new chessboard interface is on its way. It will be based on the latest Microsoft Technology called Silverlight / WPF.
We also talked about the parsing algorithm. It happens from time to time that users send me emails that CPT doesn't accept the right move. This happens only if incorrect PGN files have been imported. In general these are ambigous moves where e.g. one knight can't move because of a pin. In this case the correct notation is e.g. Nf6 and not Ngf6. Future versions will be more tolerant.
Then we played some chess games:
Unfortunately Gregory improved over the last three years and he beat me quite a few times. It looks like his regular chess sessions with Cafemilate pay-off!
At the end we talked about position transpositions and performance improvements for really huuuuge databases. Right now CPT only recognizes transpositions of the same color for a given position. We plan to support all kind of mirrored transpositions (e.g. black vs. white to move, but also left and right side). This feature is planned to be implemented for CPT 4 as it requires a different storage format of a position and I want to have this right from the beginning of this new age of CPT versions.
A pity that he already had to leave to continue his journey. I wish we could more easily see each others.
Stefan