Would be pretty cool to have conditional moves support logical operations. An example being: if my opponent does NOT play P10 OR O11, play R13. Just a thought I had while playing recently. Really enjoy the conditional moves feature as is, especially for joseki patterns.
Separately, I am a developer myself and was wondering if there is any guidance surrounding contributing code to the site, or if such volunteer PRs are even allowed.
I wonder how advanced one might want to make this logical operator language and how large of a premade file might be allowed.
Could this language be Turing complete, or, even further, convenient enough to practically implement some algorithms, like say processing the board position through a neural network and some search loops making use of that? And could the file be allowed to be large and descriptive enough to specify the parameters of a pretrained network?
I’m just playing devil’s advocate here. I doubt any such feature would actually make that practical, and, of course, it should be prohibited by the rules against AI assistance.
However, I think there is some gray area between the current implementation of the conditional moves and more advanced versions, as suggested earlier in this thread, that I think is worth discussing, in order to figure out where it makes sense to draw the line.
@Vsotvep suggested the use of premade files that could contain fuseki/joseki books. How far should one go with that?
For example, AlphaGo Teach’s entire fuseki book is available online and can be download as an SGF file, see:
Would it be appropriate for one to make and use a premade conditional moves fuseki book based on that?