Ouch … this is tough. Because, yeah, just having cower, poser, loner, hover, joker and boxer (each with different 1-3 letters) means you could just be hosed by randomness. But on the other hand it feels really cowardly to try a word like WISPY to eliminate that many more words because you aren’t trying for that 2-guess brass ring!
The conflict of Wordle strategies (minimizing average guesses vs actually getting things in 6 guesses) is fascinating to me.
As for the project … @DrkStarr had some cool transcripts which got me untracked to do some very interesting stuff for Shuffling Around. I managed to hack Zarf’s regtest scripts so they could detect bad punctuation. Inform’s line break handling causes you to do something like
say "That worked really [if yay is true]well[else]badly[end if]."
And most of the time you remember to, but sometimes you get 1 of the 2 below:
say "That worked really [if yay is true]well[else]badly[end if]"
say "That worked really [if yay is true]well.[else]badly[end if]."
One gives no period and the other gives 2. Sometimes a stray period appears on one line as well!
So I was able to check this. The interesting thing was, I got some false flags, but they also turned up things to tune up. I also managed to run something to test HINT (OBJECT) on all objects to make sure things were handled sanely. This was a small holy grail, and often just having transcripts gives me a boost to try new technical stuff, and the new technical stuff, while it doesn’t show everything’s perfect, gives me confidence at least some things are going right. There weren’t a ton of bad/obvious things, but enough was in the “I’d never have seen that” camp, I’m glad I did it.
I’m typing this as tests are running. Sometimes I run and re-run a test and I finally think “this time it’ll pass” and it doesn’t quite … but it’s been good to learn how not to feel helpless and not anticipate the test results too much (and even write a script to peek in and be able to scrub a test that misfired badly,) and how to prep the next thing I want to do without burning out. And there’s one more thing–the tests are pretty intensive on my Linux emulator, and if I busy my Windows desktop up, it slows the tests down. So it’s been a good way to motivate myself to get up and stop staring at the screen, which has been a pitfall during COVID.
Perhaps the most frustrating and amusing part of the whole test process was this: I finally got around to testing a maze, which is optional. I didn’t want to do so in detail, but I figured I should. Of course, my test script got lost in the maze several times. There were “only” 25 rooms, but even so, there were enough dead ends I failed to backtrack through that it took a lot of energy to (finally) get things working. This was a “nice” reminder (not that I needed one) to avoid putting mazes in any future games, as the pain I experience–well, I don’t want to pass it on to anyone kind enough to look at my stuff.
Overall I’ve covered a lot of things I didn’t expect to, stuff I would’ve said “no way I could do that, but it’d be neat” back in 2013 or so. So I really feel ready to move on. It’s a bit scary, in a way, as this project was more a part of my identity than I thought. But also, I know I have stuff I can and want to move on to.