grilops VS puzzlord

My open-source project https://github.com/obijywk/grilops (excuse the shameless plug) can help when creating Nikoli-style grid logic puzzles, and we used it during the development of Ents, Resolution, and Missing Pieces (and to check uniqueness of Digital Gaming solutions). The constraint solver library it depends on, https://github.com/Z3Prover/z3 from Microsoft Research, is also very useful on its own, and we used it to help with the creation of Art Gallery and Global Shipping Crisis. Dennis Yurichev's SAT/SMT By Example is an extensive resource for learning how to use these kinds of tools to solve all sorts of problems, including puzzle solving.

We wrote the metapuzzles first, of course. IIRC I believe A Man of Letters was the first of the metapuzzles that we actually used to be written and tested (although that's before we started using https://github.com/galacticpuzzlehunt/puzzlord to manage puzzle lifecycle, so I don't have the exact data in front of me). We were also working on a meta version of Resolution before we ultimately decided to rebuild it as a regular puzzle instead of a metapuzzle (and this was true of Art Gallery as well). The first regular puzzle to enter testsolving was A Curious Pairing, and the first regular puzzle to complete testsolving was Institutional Knowledge. Magic Words was the last regular puzzle we completed (and we only finished testing it on May 11 - a bit too close to release date for comfort!). Using number of Puzzlord comments as a proxy metric for "went through the most changes", Art Gallery had the most (269), followed by Into the Words (255) and then Magic Words (250).

There's a few hours of reading in here, but even just skipping around through the post has given me a ton of new ideas for my hunts. If you're confident using Github, you can even download their program used for editing and testing puzzles.