I fell for this before. Wasted a few weeks' free time adding thumbnails to the GTK file chooser, only to find out that the library itself has bugs which they refused to fix that make it difficult to do. Now I just have a patched version installed locally. The only way to improve these projects is to change how they're managed, which you can't do with a patch.
There is thumbnails support in the GTK4 file chooser now. That probably required shitloads of fixes in GTK, but now it works.
The way to change projects is to write good code and work together with the other devs that made something you wanted to contribute to and then improve it together.
I fundamentally disagree. The problem isn't that there were no thumbnails, but that it took decades to implement them. The cause of that problem can't seriously have been that people weren't programming hard enough. I think lots of code was likely written in that time. If we assume, as you seem to, that the issue was not enough code being written, then GTK would need to increase its count of contributors by 120 times to ship that feature in the span of two months. Your implication here is that it takes thousands of people multiple months to program a file chooser with thumbnails. I think one person should be able to do it in that time, and that the problem was something other than writing code. What has changed about the project to guarantee that the next important feature won't take 20 years to implement?
There were no thumbnails because no one wrote the necessary infrastructure before some one actually did. One person did most of the work and there were no other problems than actually writing code.
Okay, so what were all the other people doing? If it's one person's worth of work (as I suggested), and there was at least one person working on GIMP, why did they not do it? The fact that it was easy to do doesn't explain why it didn't happen. If anything, it makes it more mysterious. Unless you are genuinely telling me that there wasn't a single person working continuously on GTK for 20 years. It seems to me that they lack a process for deciding something needs to happen and then doing that thing.
The people who are doing the work, or a paying for the work, on GTK think that other tasks are more important or fun to do than the tasks you wish should be done. You might think that they are wrong and that other features are more important than a modern accessibility stack or whatever else the GTK devs have been doing, but the only way to change that is to actually engage in that project.
This is far from a personal opinion. It's probably the number one complaint about GTK, and it affects almost all software that runs on Linux since the GTK file chooser is often the default system file chooser. Almost all software opens a file at some point. Web browsers, chat clients, etc. Photos on user's computers rarely have meaningful names because they are often taken with a camera and the default image name isn't changed. This means that there has been no ergonomic way to open an image file on Linux for the past 20 years, and there still may not be for software that uses older versions of GTK. The accepted method is to open the system file browser and drag a file from that over into the application you need the file in. This has had an enormous cost to the reputation of all open source software.
