Quoll Work Thread, 2024-09-26 - 2024-09-26 14:58
I don’t intend this to be an actual thread like the last one because I’m more wrapping up for the day by re-running all the tests. I did the work before I made this post, as I sometimes do when it’s pretty small. But I do still like talking about it, so here we go.
Today was wrapping up the last few fixes and improvements I wanted to do with the Smarter Parser integration.
My first real commit today was updating the default noun example to be “BANANA” instead of “FLOWER” because it is a
funnier word. But I also added my own version of the [get noun example] phrase to skip over software and Things that
are part of other Things, such as the screen and keyboard for the phone. This is mainly because while these are
definitely implemented as Things in the model world, they don’t strike me as something I want the Playerr to really
know about because I am usually not requiring them to interact with them.
Then I made some fixes to where typing like “NOVICE ON” when novice mode was already enabled gave a message about the
verb being unrecognized. This is a result of how Aaron Reed wrote the Understand clauses, with essentially the command
only being understood when it would do something. So simple enough to add some Understand [...] as a mistake clauses:
Understand "novice on" as a mistake ("[as the parser]Novice mode has already been enabled.[as normal]") when novice mode
enabled is true.
Understand "novice off" as a mistake ("[as the parser]Novice mode is already disabled.[as normal]") when novice mode
enabled is false.
Understand "correct on" as a mistake ("[as the parser]Auto-correction has already been enabled.[as normal]") when
corrections enabled is true.
Understand "correct off" as a mistake ("[as the parser]Auto-correction is already disabled.[as normal]") when
corrections enabled is false.
The same commit as that also updated a Smarter Parser rule around catching the Playerr making assertions to account for the possibility of singular “they”. Basically, the default rule would catch “SHE IS WEIRD” but not “THEY ARE WEIRD”. I honestly expect lots of Playerrs to refer to Glitch using “they” despite that being a misgendering (Glitch is cool with it; it’s OK).
The last little bit was adding more explicit hints to EXAMINE Glitch’s notes in the apartment when they haven’t been examined yet. I also wanted to add a more explicit offer of a tutorial when the Playerr is in novice mode, but that would require me to have to update the tests to handle the prompt for the tutorial, and that’s more than I want to do right now.
But with that, I think I am at a good point with my Smarter Parser integration for now. I do think based on what I’ve noticed with playthroughs that it may behoove me to discard the default rules and make my own rules. There were several cases in our playthrough where I was very surprised which Smarter Parser message came up, and while it works out, I do want these messages to feel more relevant than that.
And since I’m at a good point where I think I can move on, I’m re-running all the tests just so I can get all the changes. I don’t actually expect changes for the Smarter Parser stuff, but there will be a lot of cases with the notes. But my tests take some time to run so that will probably do it for me today.
(Although, quite worryingly, the Inform 7 editor seems to have issues saving my Skein while a thread is open. And even after closing it fails to save unless I wait for a bit. There must be some processing that is taking a bit long because I’m straining the editor with my very large Skein. I’m just going to carry on until it becomes a huge blocker.)