Thank you for providing links to true third party clients and not just webview wrappers and mods for the first party client.
Privacy isn't the only reason to use free software. Some people have contacts on Discord that they don't want to lose touch with. Of course, like all proprietary silo networks, Discord is best avoided if possible.
If you're hosting it yourself you're still legally liable but at least you get a chance to defend yourself instead of being betrayed by your host with no recourse.
Anyway, emulation isn't illegal so it's not even clear these developers did anything wrong, but proprietary web 2.0 services shoot first and ask questions never because that's the best way for them to cover their asses legally.
Because even some FOSS developers value convenience over freedom and privacy, unfortunately. Those of us who've spent decades warning about this get shouted down and told we're being too puritan or ideological and that the proprietary thing just works and everyone uses it already so we're wasting our time.
I don't think the ffmpeg maintainer is complaining that Microsoft is using ffmpeg, rather that they are opening "high priority" bug reports based on customer complaints. This might be a high priority problem for Microsoft but that does not make it so for ffmpeg.
The license allows Microsoft to use ffmpeg but they aren't entitled to demand free labor from the project. Really, no one is entitled to do so, but Microsoft being a large company who can definitely afford to put money or talent on the problem makes it only that much more egregious.
edit: I would note that asking for help or reporting a bug is usually welcome, the problematic part is demanding help because it's a high priority issue for YOUR customers.
Very concerning misinformation in this thread. Open source does in fact mean more than "can look at the source code." The open source definition closely parallels the free software definition, in fact.
I don't like the terms open source, FOSS, or FLOSS precisely because of this misconception.
Fauxpen source licenses such as this are the answer to the wrong question.
"Other people making money with my stuff" was never a problem in the software-freedom community. Whether this means "selling my stuff" or "using my stuff in a commercial setting" ("commercial use" restrictions are confusing in this way). In the free-software world we just accept that our work belongs to the community and the community can use it in ways we don't approve of.
(Edit: Likewise, it has never been an issue to sell copies of free software, although I should point out the very nature of software freedom makes it more difficult to guarantee a revenue stream in this way)
Rather, this is a symptom of the proprietary software world's reaction to free software and co-option of it (in the form of the open source movement). Tom Preston-Werner, founder of GitHub, opined that proprietary software companies should open source almost everything - "almost everything" being anything that does not "represent business value." In other words, open source cost centers but keep profit centers proprietary. Ideally, these companies would cooperate on widely used components (and some do!), but practically they spend as little as possible because capitalism. This is also why we see so many projects turning fauxpen source lately; these companies imagined they were developing cost centers and then realized they could be profit centers instead.
What was (and still is) a problem is people making proprietary derivatives of free software, and copyleft is the solution to that. If you want to extract license fees from proprietary software developers you can dual-license under a strong copyleft like (A)GPL for the free software community and sell proprietary licenses. Believe it or not, Stallman explicitly does not object to this - mainly because, if selling GPL exceptions to enable proprietary development is wrong, then releasing under a permissive license must also be wrong because that also enables proprietary development.
xkcd 743 moment
Traditional GNU/Linux distributions (as well as F-Droid) are not "app stores" even though they are superficially similar. Traditional distributions are maintained and curated by the community, and serve the interests of users first and software developers second, whereas an "app store" has minimal curation and serves the needs of software developers first and users second.
I point this out because there's an annoying meme that traditional distributions are obsoleted by the "app store" model. I don't think that's the case. "Verification" is essential for an app store but pointless for a distribution.
As I understand it, the archive repo hosts old versions of apps that are in the main repo, it does not itself host old apps.
A tool with fewer features that is harder to use is by definition an inferior tool.
That's only your opinion, not an objective truth, and I only partially agree with it. Having the most features is not as important as having just the right set of features, and there are anti-features to consider as well. Feature creep can actually impact the usability of a tool, so these two criteria are sometimes in contradiction.
Ease of use is subjective and depends on the user, because users' needs, ability, tastes, and concerns differ. Of course, I don't think anyone deliberately chooses a tool because it is hard to use.
I don't agree that freeness is purely an ideological concern. I don't think a tool that works against me, or imposes arbitrary restrictions on me is a good tool by any measure. A good tool doesn't enshittify, or spy on its user, or refuse to work for arbitrary reasons. If a tool doesn't work and you are legally not allowed to fix it (as in the printer which inspired the movement in the 1980s), it's not a good tool. If a tool punishes you for something you didn't even do (as BitKeeper did to the Linux developers) it's not a good tool, even if it has the right features.
I don't tell you that your opinion is wrong, only that I don't agree with it. We are told our concerns are invalid and don't matter.
Beeper is proprietary but their Matrix bridges are all free software, as far as I am aware. I think you can also use a free software client with Beeper's service (that's what people have said in the last Beeper thread I've seen).