freedesktop - Apr 2017 - [fdo] Ideas for improving current situation on Linux/UNIX desktops

Hello everyone,

Here are some general thoughts as an end user. I really apreciate an 
initiative such as the freedesktop’s that promotes collaborating amongst 
desktop environments (DE). I saw recently some good initiatives going to 
this direction with the merge of Razor-Qt and LXDE to create LXQt, after 
their respective teams found out they have a similar vision, and the 
collaboration between the LXQt and KDE teams that allowed LXQt to be 
used with kwin. It’s great to see a collaboration between 2 projects 
with a very different vision, but I believe it is only the beginning of 
what could be done. As far as I understand, freedesktop puts especially 
the focus on creating a common backend and defining standards and 
protocols to ease interoperability between DE and to reduce dupplication 
of code.

My question is, why not go further with these ideas to improve the user 
experience and reduce dupplication of code, not only on the DE backend 
level but also at the application level (pdf viewer, text editor, 
music/video player, etc) ? Currently DE keep reinventing the wheel and 
create their own applications in order to give an unified desktop 
experience to their users. Personnally I use the application that better 
fits my needs, is it the one written for gnome, KDE or anything else. 
But the lack of unity that results from that in my desktop experience 
still sucks and installing KDE or Gnome dependencies because of one 
application is overkill.

I find it also puzzling to find between 10 applications designed for 
doing the same thing, which one does it better for me. I believe this 
situation to be a big loss of time and energy that would better be used 
for some more valuable work. Doing its own things looks like more a 
closed source mentality than an open source one. Open source really 
shines when people have a collaborative mentality. On the other hand, 
having developer teams with different visions and targeted users and 
hardware is a real richness, and it’s necessarly for innovation, but 
they should ideally collaborate together as long as it doesn’t conflict 
to their vision. I mean, if 2 teams want to do the same thing but could 
only agree about 80% of what the features should be, then they should 
better do the 80% together and do the remaining 20% alone instead of 
doing everything alone.

For me it seems that a good solution for improving the situation with 
applications on the Linux/UNIX desktops would be to have one application 
shared by several DE and structured in a modular way with a core part 
independant from any desktop environnment or GUI toolkit, this being the 
lowest common denominator, and a DE and GUI toolkit dependant part that 
will give the expected user experience and look&feel for the given DE. 
Would it be any technical or performance issues which such a solution ? 
Or is it not happening merely because the DE teams aren’t ready for 
collaborating that close to each other ? Ok there will probably be cases 
when collaborating wouldn’t be possible because of divergence of 
visions, but at least I believe that collaborating should be the first 
instinct and doing its own things should be a last resort.

I am far to be an expert on the subject, but I have these questions in 
my mind since a while and I would like to go deeper in that matter.

Cheers,
Jonas