I’m back, and as promised I’ve been putting some more work into the Salta game. The latest version is now up on its own page, which will house update links and the comment thread from here on. Thanks again to everybody who has so far spent time with my ramshackle pre-alpha game.
Release notes for v. 0.3.1:
- Statically linked to the MSVC runtime dll, which should prevent a known source of “Error defining an external function” failures on pre-Windows Vista operating systems.
- Switched the AI from pathfinding back to maximizing score when the opponent seems to be getting near to ending the game, to avoid embarking on a short-term setback for long-term gains that will never come.
- Rewrote the tutorial speech balloons and the credits to use display methods that should work regardless of window placement and screen resolution. Thanks, kind of, for the patience of everybody for whom the old method could not possibly have worked well.
- Fixed ESC key and Window close button to get confirmation before quitting the game.
- Implemented Undo. This will unwind to just before the most recent human player’s move, so that’s always two moves in the current configuration.
- Added a more specific tutorial message when you try to jump one of your own pieces.
- Added a “Hint” button. This will run the AI on your current board position, and tell you what the computer would do in your place.
- Several under-the-hood changes that players won’t necessarily perceive. E.g.
- Consolidated all the buttons into one object type
- Moved the .dll interface over into Game Maker extension packages (which are more work for me, but at least aren’t officially deprecated, so may work better).
- Bug emails now automatically fill in a subject line with the version number of the executable
Unrelated to the current release, I also did some testing to see how many path searches were turning up a route at all. I needed to check this to see whether A* search was likely to be an improvement over Djikstra’s method, since they should take similar time to tell that there is not a path. But there is a path of some sort in about 40% of searches, which is not as bad as I would have guessed. I know, I know, this is probably fascinating to all the non-AI programmers out there.