[bulatb]

Quality is what we call the difference between what succeeds and what we wish succeeded. It's a measure of how closely the opinion of the person saying "quality" aligns with what is actually rewarded by the fitness function that objectively exists. The only thing you've learned when you hear "quality" is what the speaker thinks is good, not anything about the subject.

Quality does not exist until a (person, preference, subject) pairing happens. It's not real. It's literally not a thing. It's just a name for mixing up your is and ought.

When someone saying "quality" can point to the specific property that makes the subject quality, they're almost always pointing to what is, from the perspective of the fitness function, inefficiency, a waste of energy and effort, more spent than the system rewards.

Anyone who spends a lot of effort building skills to get particular results is going to believe those results are important, at least a necessary part of success, if not the whole definition. When people who did not put in the effort keep succeeding without those results, "quality" becomes the cope. It's shoddy work. They did it wrong. They cheated somehow.

And yet they succeeded.

Because "quality" is what a person says when is and ought don't match, a fact about the speaker, not the subject.

[nomel]

I agree.

I think the only time quality might be objective is when there's a planned "ought", like a required feature that needs to be implemented later.

From that perspective, the "quality" could be measured in how easy it will be to implement that, in a "total effort" sense (subjective I suppose), where "low quality" might require a whole rewrite of big sections, where "high quality" would have the required stuffs in place (maybe abstractions, or maybe even placed where spaghetti would fit nicely) to make it "easy".

[bulatb]

Yay! I've argued this a few times here and it's been, uh, perceived as low quality.

I think "objective quality" can always be re-framed in terms of fitness. "Quality" is just not anything the fitness function measures, so it can't.

It's hard to really make the case because of course there's something real that people mean, colloquially, and it's totally valid. Lots of things objectively have properties most people would agree show quality. That's where it is, it's right there!

But all those properties can be described without the use of "quality" by just reciting every fact about the thing. The only thing that will be left, the only thing you can't communicate that way, is what the person thinks about those facts.

[nomel]

> .. perceived as low quality.

What you're saying requires someone looks past some existing categorizations that they've formed. I suspect you're not as vulnerable to categorization as the usual person, so this probably happens often. ;)