Doctormo made a tought-provoking video, which I think is worth listening to, and pondering about.
Here are my thoughts on it:
Indeed, as a user, you cannot expect any Free software project to continue to indefinitely serve your needs, unless you are somehow a major contributor to it.
A good example is MuseScore, a program for typesetting music (which nowadays I mostly use for my own needs, but I did do it semi-professionally for many years, mostly back in the day when I was using proprietary software like Finale). With version 4, MuseScore became completely unusable for me. As it became more and more popular and gained more and more windoze users, it has focused away from Linux, and removed essential Linux features I was depending on. It also added some proprietary add-ons and spyware, which I simply refuse to run. So, I had two options: use this crippled version of MuseScore, stick to the previous version (version 3), or switch to something else. Knowing that staying on version 3 is at best a temporary solution (especially on Arch), I decided to immediately switch to Lilypond, which quickly proved to be a great decision and I’m sure it will work out in the long run. However, most of the times such an alternative is not available.
I think that is one of the reasons Free software is not as popular among professionals. There is this perception of unreliability in the long run, and professionals really do not like to change their workflow. As a Free software user, I’ve grown accustomed to dramatically changing my workflow, more often than I’d like to. On the other hand, there certainly are benefits to being often kicked out of your comfort zone, but it’s not pleasant.
I think Free software should be Linux/BSD focused. Now, it’s very easy to dismiss that statement as bias, based on my passionate hatred for anything Micro$oft, but I think it makes sense even without that bias of mine. Focusing solely on Linux/BSD means having a very small user base, but also a very devoted one. Windoze support should be offered as a third-party fork, but not be something the official project should devote any resources towards. In principle, it is very easy to make Linux software run natively on windoze (in practice, it there may be problems and new bugs), pretty much all of the dependencies are available there. Windoze software, on the other hand, typically has proprietary dependencies which simply do not exist for Linux.
When I see a Free software project that does not cater to windoze users at all (my favourite example is GCC), I immediately consider it to be more reliable and trustworthy – it shows that the developers take Free in Free Software seriously; it’s not just about being Gratis software, or just about sharing source code, but having an ideological alignment as well. I have immense respect for software developers who do not even have a windoze machine to test their software on, and say to people demanding a windoze-version: “it’s Free, if you want to make it yourself, I won’t stop you, but I want nothing to do with it!”.