I think part of this discussion needs to be that Firefox, while open source, doesn't provide footing for the community to build browser forks the way Chromium does. This sort of open usage, ironically, feels very fitting to the spirit of Firefox, but they're not even close to Chrom(e|ium). I've been building a browser full time the past year+ (synth.app) and FF core wasn't really even an option. They've done very lit…
That's a real missed opportunity. Firefox would have made a fantastic embedded webview, and could have plausibly become the core of Electron. Back when webviews were IE, with atrocious Javascript support and all the usual IE quirks, Firefox was a superior product and interconnecting with other software is very in line with the open source ethos.
A lot of GUI based programs should think of implementing parts of it as a library others can use. I remember early Git implementations in IDEs where they just ran git commands in the background and parsed the command line output. Imagine if the same had been done with SQLite, would it still be as popular today?
I don't have much experience with creating desktop software, but it seems to me that separating the GUI from the "engine" would bring other benefits too.
I have often looked at OSS and wanted to use parts of it myself. An email client might have a really good mail parser that handles encoding, attachments etc. But decoupling it from the rest of the program would take too much time.