For the last two days, I’ve been working with someone on a program.
This came about because they run a rather famous puzzle website, full of puzzles of tons of different types that they generate via computer. That site has a puzzle type that is close, but not quite, like one of my favorite Nikoli types, Ripple Effect. And their site has a bit that says “Don’t see your favorite puzzle type? Let me know!”
So I did. I even sent them photos from a puzzle book I just got this past Saturday, that I’m in the middle of solving, which happens to have some Ripple Effects in it. They said that they’d take a look at it, and honestly I thought that that would be the end of the whole deal.
Then they contacted me, asking: are you willing to transcribe some puzzles, so that they can be used as test cases for the solver?
Here’s a (puzzle) thing: I like to transcribe puzzles. I’ve done so for the tiniest fraction of the number that I’ve solved over the years, although if I had infinite time and energy I would put every single puzzle I’ve ever solved in a computer-readable format. That probably sounds really stupid, but I feel that puzzles are a fascinating form of entertainment, and one we treat entirely too ephemerally, tossing the magazines once we’re done with them. I fully realize that I have a biased view here, but is it so crazy to think that in 100 years someone wouldn’t want to try and solve a full issue of, say, Nanpure Fan in the same way that I sometimes enjoy leafing through old advertisements? It’s not like the type of puzzles I do (i.e. non-crossword-y ones) have an expiration date. That sudoku will solve just as well in 2118 as it does here in 2018.
Anyhow, of course I said yes, and set about immediately to transcription. Well, no. It turned out that the format that was being suggested was actually a bit of a pain to write by hand, and I thought of a way to make it easier, so the website’s author sent me the code he was using and I ended up banging on it to suit my will.
As of today we’re sharing our work in a Github repository; I’ve mostly done transcription (of course) and some code cleanup, and he’s been working on the actual solver and friends, since this is all based on code he’s been the master of for years.
If the generated Ripple Effect puzzles actually get published, I’ll link them here, but even if not it’s been nice to work on some code for the first time since I retired. And it’s been nice to collaborate, to bounce ideas back and forth on how to do a thing. That’s definitely something I miss from the job.
And, if nothing else, it got me to preserve at least a few more puzzles from a couple of books I have. Probably no one else cares, but perhaps one day they can be used to reconstruct a puzzle book, to be solved by our ancestors who may have never touched a physical magazine at all, to be experienced all over again.