It's even simpler than that. Who wants to work on changing a UI that they themselves are comfortable with? It's work with - best case - no benefit in the eyes of the developer. Worse, you likely will break workflows - including your own. Sexyness doesn't need to factor in at all.

And do note the nice uncanny valley of no change: in order to improve the UI, you need to understand the program well. In order for that to happen, you need to be proficient with it. In which case, you know how to use the UI and are more likely blind to its downsides.

One more thing: some people have no eye for aesthetics. I’ve seen (smart) people defend a piece of code, and when I show them a refactored version, they say “that’s literally the exact same thing I wrote.”

No it’s not, and yes it is, in the sense that’s a^2 + 2ab + c^2 is not the same thing as (a+b)^2, but they (should and ideally must) evaluate to the same thing. I say ideally because only in the case of using infinite precision arithmetic can you guarantee this for all values of a,b,c.

People blind to these patterns cannot see how to simplify things, whether math or code or UIs.

Was the use of c^2 intentional?

Lol no.

I think developers and most users simply don't even notice the problems.

When I say "I can't use program X because the UI/UX sucks, is there any alternative?" people come out and say "I see no problem with program X, I use it everyday and it rocks"

Vim rocks, yet a lot of people can't use it. Once you state "X sucks", someone would object.

You've got strange conclusion out of it.

Yes. And good UI design requires a look of work and effort to get it right. You need many skills, or a good team to do it and that's rare in FOSS.

In my opinion, improving user experience isn’t something that is easily done piecemeal. It’s one of those things that can require architectural changes to get right, and so it’s difficult to improve on existing projects. Fixing bugs and making things faster is “easier” in a sense because the end goal is obvious and you can target specific portions of code to make immediate improvements. Improving UX can mean making many architectural changes before being able to reap any fruit from your labor. I say this as someone who has worked on a lot of apps. I can’t imagine the difficulty of making several pull requests and getting them accepted without any clear output in the product itself.

If you're suggesting that this is done out of selfishness or intellectual fervor then I don't think that's true. I wouldn't say they "prefer" it but it's certainly easier and more economical for a skilled person to contribute to those tools because the developers and the users are the same group of people.

Once you decide you need to do user testing and hardware testing, the cost shoots up. You need to start procuring hardware and finding a place for it to live so automated tests can run on it all the time. That costs money. You need to bring in a lot more testers which requires project management and/or volunteer coordinators. Usually that costs money too. Everything you do needs to go through more stages of design review which also requires extra coordination and project management. These are real things you can have but someone's got to foot the bill.

> Once you decide you need to do user testing and hardware testing, the cost shoots up.

It is not so bad.

Take three people who never used given software, ask them to do the most basic tasks. And fix the most common problems.

Your (and mine) software is much harder to use than you expect.

You do not need UI/UX people, massive scale testing to fix low hanging fruit.

Just my experience, three people is not a big enough sample size for most projects. You will get very biased opinions if you go with that. That may be ok if you have a very niche project and you only ask some people in that niche.

> You will get very biased opinions [with 3 testers]

Your anecdote is atypical.

https://www.nngroup.com/articles/why-you-only-need-to-test-w...

Or when you have obvious problems!

Take it with pile of salt, I am not some big expert. I just did it with some projects. But it seems to me that in nearly all cases there are glaring issues that would be detected with any amount of testing at all.