Ericbsd: All of your points are valid, of course. I just wanted to express my take on the direction that GNOME is taking, tying more closely with Systemd and all that stuff. If the TrueOS team decided that it's actually more feasible to write their own DE instead of having to keep patching KDE, that's a pretty meaningful thing. And as far as I know, GNOME is much worse compared to KDE in this regard. This might lead to GNOME³ stopping to work on anything not Linux/Systemd some time in the future and thus it would have to be dropped again in GhostBSD. But these are just my thoughts and doubts and if you think that there's value of adding GNOME³ to GhostBSD again, I'm of course fine with it.
ericbsd wrote:
Now that is cleared out of the way there is three way to look at it.
- We can continue waiting on the FreeBSD Gnome team, ports Gnome, MATE, XFCE and other software.
- Join our effort with the FreeBSD Gnome team to help to port MATE and all other DE and to try to fix problem upstream.
- Starting a new GTK DE for FreeBSD and focus only at that DE.
You know FreeBSD's GNOME team and can probably guess how it will act in the future. I only check the quaterly status reports from the foundation, always hoping that there's some news from the GNOME team there. Unfortunately to the outside world (which includes me) it looks like things are going rather slow there. Unless you know that some more people joined in, option 1 is probably not the best thing, even though it's of course the easiest one.
The second option sounds right to me. You made a big stop forward in porting MATE 1.18 over. I don't know how it's supposed to work, but will the GNOME team accept it and help testing or polishing it? There have been ports for MATE 1.14 and MATE 1.16 around but they never made their way to the ports tree. Why's that? Were they lacking in quality? Did nobody really bother to make a release? Understanding things like that could surely help to work together.
And about the last point in option 2: That would be the right thing to do, of course. Are both GNOME and MATE accepting patches to make things work on FreeBSD? Sometimes there are projects after all that explicitly "don't care about anything other than Linux" which would of course be a problem for us. The prefered situation would of course be if us BSD users would manage to upstream all the patches and then get track the development versions of both projects so that we hit breakage before a release ships and don't have to patch afterwards. Well, actually the best thing would be if we could somehow provide a FreeBSD CI machine or something so the upstream devs notice right after their commit when they broke FreeBSD support.
Most people that know me, they know that I would like to make a GTK DE for and made on FreeBSD/GhostBSD, because there is only one DE that is made on FreeBSD/TrueOS and it is Lumina. It is maybe not the best DE I ever use, but it works better than all other DE on FreeBSD.
There is more GTK DE than QT DE and there is one DE that I have use that works better that all other DE on FreeBSD that is XFCE and it is partialy because they keep it simple.
Now I do prefer MATE look, but don't like the gnome complexity they have keep. I ensure you that if we could have the supports needed to create a DE I would start it right away.
I feel the same here: A native DE for FreeBSD (or one that aims to be ultra-portable) is definitely the superior approach. Well, to be yet a little be more utopic, a BSD licensed and *BSD aware graphical toolkit would also be dreamy. But obviously the BSD community doesn't really have the resources to do that (or rather: Other priorities). Lumina was an awesome idea and for some month I tried very hard to love it. Unfortunately it's Qt which I don't like a lot and it doesn't have the right feeling for me. So at least IMO there's room for another BSD desktop besides Lumina. Maybe one day the time will come.
Here's one more thing that crossed my mind: We have LXDE which is a pretty nice little desktop. It's currently maintained but the maintainers announced that it will be dropped in the future in favor of LXQt, their Qt port. Maybe when that happens it might be feasible to pick it up and continue that, gearing it more towards a desktop for all *nix systems? The big problem is definitely that it's still GTK+2 and would have to be ported to GTK+3 in the future if it is to be kept alive. I would imagine that there are some devs which don't like the GTK version dying who would be interested in keeping it alive. But again, this is just a thought.