We don't have dedicated QA at my job, but personally I'm thrilled when a coworker finds a bug in my code. It's much more preferable to the customer finding it
The only problem with that is the vast number of declensions. Sure they're not as wildly divergent as, say Latin or Ancient Greek, and there's no gender, but because of all the cases there's a lot of subtle variations to remember
> A circadian clock must be self-sustained and internally driven, as the 20-hour cycle of the jellies’ spawning is. It must also be regulated by an environmental stimulus such as light; while the jellies’ spawning clock can run on a 20-hour cycle under persistent light in the lab, in nature it resets every day.
It's not just media files. I'm forced to use Windows 11 on my work PC, and I had to disable the new shell extensions to make the file explorer usable again. It's noticeably faster without the new UI.
According to the wide definition of grammar, everyone uses some grammar, but nevertheless it's not a category error to say "*this sentence got grammar mistake".
When we speak about the grammar for a language/dialect we imply a prescribed "correct" grammar for a particular community of speakers.
> speak about the grammar for a language/dialect we imply a prescribed "correct" grammar
Not prescribed, rather observed. At least in English where there is no language authority and dictionaries present usage.
The situation is even stranger in Norway where there is a prescribed form but where dialects have essentially equal rights so that the prescription really applies only to formal written Norwegian.
From what I experienced, "proper" English grammar is absolutely prescribed at schools, and "poor grammar" will consistently and very predictably get you point deductions, regardless of whether linguists would accept your grammar as valid.
I've been taking a look at my sound change applier again. I've been trying to add syllable detection to it so that I can match directly on syllable boundaries instead of having to match on them explicitly in my sound change rules.
So I started by adding the ability to define syllable structure in the rules file, then I tried running the syllable rule through the same compiler I used for the regular sound change rules. It ended up being even slower than I was anticipating, so I decided to skip the NFA to DFA conversion step and wrote a backtracking NFA runner. This worked _okay_, but if the syllable rule isn't able to fully match a word it ends up backtracking forever, and I never managed to figure out how to fix that.
Last year I read a post about parser combinators and I decided to rewrite the syllable detector. I finished the rewrite and then ran into an error and gave up. This last weekend I revisited it and it turned out it was just user error again; my syllable definition rule had a mistake, but thankfully the error was a lot easier to fix with the new design. Now it emits a warning, and I'm rewriting my sample sound changes rules to use the new boundary markers and hammering out any issues, which are a lot less than I was afraid of.
I'm thinking about rewriting the sound change rule compiler to use the same combinators I did for the syllable rules, but it would be kind of a shame after all the work I put into the DFA compiler lol
reply