Larry Hosken: New: Crossword Compiler Noob Diary

Unsurprisingly, creating mediocre crossword puzzles is easy but creating good crossword puzzles is hard. Mind you, I don't feel pressured to create great crossword puzzles. For puzzlehunts, I only need to create acceptable gimmicky crossword puzzles. Solvers might roll their eyes when they see a puzzle that uses both TSAR and CZAR, but maybe they'll forget all about that when they notice that the Os form the shape of the Big Dipper or whatever. But so far, I'm not even making acceptable puzzles.

I think my next hurdle is compiling a good word list. Crossword Compiler comes with a pretty darned good set of lists! But it seems like it could be better. It doesn't know about Hank Aaron. Hank Aaron's kind of a big deal, right? I don't know baseball, but I've heard of Hank Aaron. I tried throwing in the Nutrimatic 100k phrases list. As you recall, Nutrimatic is a smart word-finding tool that normally uses a big pile of data from wikipedia. That knows about Hank Aaron. Unfortunately for this purpose, it also knows about phrases like "of this" and "from his" and "category list of left handed accordionists". OK, I made that last one up. I think. I'm not sure what exactly would happen if I made a crossword puzzle full of stuff like "of this" and "from his", but I suspect it involves Tyler Hinman bringing back the #badpuzzles Twitter hashtag.

I took one step in a good direction. I went back to Nutrimatic's wikipedia data. I cranked out a bunch of Nutrimatic-ranked phrases, but then filtered that list to only things that were titles of wikipedia articles. Although "of this" appears many many times in Wikipedia, there's no article about the phrase (yet). So that's a step in a good direction... but it's just a step. Kind of a stumbling, gimpy step. Consider that Wikipedia has information about John R. Adler, John R. Bolton, John R. Allen, John R. Commons, John R. Lynch, John R. Brinkley,... Uhm, you get the idea. It also has an article about a I-hadn't-heard-of-him D.J. named "John R." My system sees all of those John R. This and John R That and happily thinks "John R" is a common phrase, so it enthusiastically suggests that "John R." is a good phrase for crossword puzzles. And it's perhaps overfond of mononymous Bollywood stars. And it makes other... questionable choices.

This seems work-aroundable. I bet there are good word lists out there. I don't know where to find them. I can find Scrabble cheat-lists aplenty. I can find lists of many obscure words... words about as much fun to solve as "from his". But I bet I can eventuallyfind good word lists. Meanwhile, my tinkerer's heart enjoys an excuse to tweak better words out of Nutrimatic. Meanwhile, I jot down a blog entry so I can remember what it was like to get started. This is today, as the saying goes.