Donna, a new UCI chess engine by Michael Dvorkin

[mid]

Congrats on creating a new strong engine with a new language. It is not the easiest thing to do! When I started to write RomiChess I only knew a little basic and nothing about chess programming. So with a C primer and a simple book on chess programming I somehow managed to create RomiChess.

Thanks Michael! Your experience sounds familiar, I had certainly underestimated the complexity of building the chess engine that plays competitively. Couple of personal observations:

- The challenges seemed to be piling up as I went along (move generator, problem solver, search, UCI protocol, time management etc.). Each challenge came with its own set of complexities so keeping focus and concentration was paramount.

- Building competitive engine initially looked like a huge achievement all by itself. Now that I've got it playing it actually looks just like the first step in thousand miles journey. One can optimize the engine forever.

What I suggest is for you to rewrite your engine in either C or C++. You will do your engine justice by doing so. Also it will give you a version of your program to use as a benchmark to compare the performance of the Go language against if that is important.

I've done enough C programming and rewriting the engine in C doesn't look particularly interesting to me. I've started experimenting with Rust and Swift however, both are pretty exciting technologies. One day you might get introduced to Ms. Rusty Donna, and Ms. Donna Swift

Best,
Michael
-- http://www.dvorkin.net

[spambanane]

Do you know if this changes something about the problem of Graham?

probably yes. just try.