You are not logged in.
Pages: 1
I'm using CentOS 7 and at the end of last year there was a mammoth update. This transformed the desktop environment. Having searched the web, it seemed the developers took a, "we know what's best for you - get used to it" attitude and so I have spent the last couple of months trying to get used to it. Still, it's too disruptive with so much functionality stripped out, and I'd like to get back the previous familiar usability. Having searched the web, it seems quite a few proponents of the new system respond to such questions with, "we know what's best for you - get used to it" attitude - citing that forcing the new paradigm on everyone prevents maintaining compatibility with the previous usability paradigm from being feasible for most developers. Maybe that's true, to an extent.
From what I have come to understand, there are quiet a few components involved in this. Some articles refer to this as Gnome 3 superseding Gnome 2. And that XFCE is affected because it has dependency on GIMP libraries, of which there are GTK -3, GTK -2, GTK +2, GTK +3.
Is it possible to convince XFCE to use libraries provided by MATE as a derivative of Gnome 2, instead of using those provided by Gnome 3? Or to tweak a config file to specify explicitly whichever GTK version to use? Or both?
Offline
Hello and welcome.
I'm using CentOS 7 and at the end of last year there was a mammoth update. This transformed the desktop environment. Having searched the web, it seemed the developers took a, "we know what's best for you - get used to it" attitude and so I have spent the last couple of months trying to get used to it. Still, it's too disruptive with so much functionality stripped out, and I'd like to get back the previous familiar usability. Having searched the web, it seems quite a few proponents of the new system respond to such questions with, "we know what's best for you - get used to it" attitude - citing that forcing the new paradigm on everyone prevents maintaining compatibility with the previous usability paradigm from being feasible for most developers. Maybe that's true, to an extent.
Can you give some concrete examples of where you see this functionality that was stripped out? Of all the desktop environments out there, Xfce is one of the slowest to change, focusing instead on maintaining existing functionality. Your comment surprises me.
EDIT: I think I may have mis-interpreted your first comment. The change from CentOS 6 to CentOS 7 was a change from gnome 2 to gnome 3. I think thats what you may be referring to. If so, then yes, I agree. There was a fundamental change in the desktop environment.
From what I have come to understand, there are quiet a few components involved in this. Some articles refer to this as Gnome 3 superseding Gnome 2. And that XFCE is affected because it has dependency on GIMP libraries, of which there are GTK -3, GTK -2, GTK +2, GTK +3.
There are actually two main GTK toolkit versions: GTK2 and GTK3. There is also QT (think KDE). Xfce is still mostly GTK2 - but the next major release is going to be a port to GTK3. See: https://wiki.xfce.org/releng/4.14/roadmap.
Is it possible to convince XFCE to use libraries provided by MATE as a derivative of Gnome 2, instead of using those provided by Gnome 3? Or to tweak a config file to specify explicitly whichever GTK version to use? Or both?
Currently Xfce supports both GTK2 and GTK3. Most of the core Xfce apps are still GTK2. I believe Mate is also based on the same GTK2 libraries. Much of this, however, depends on the application that you use. If the application has been ported to GTK3, then it will use the GTK3 libraries.
My understanding is that Xfce and Mate have more in common than they are different.
Please remember to mark your thread [SOLVED] to make it easier for others to find
--- How To Ask For Help | FAQ | Developer Wiki | Community | Contribute ---
Offline
Xfce is one of the slowest to change, focusing instead on maintaining existing functionality.
Indeed a major reason why I use CentOS and XFCE is that they are feature-stable. This allows me to get on with my workflow in a familiar environment. That's why I was so surprised such a major change arrived in an update (rather than waiting for the next major version) and it has meant my workflow has frequently been interrupted this last couple of months with google searches for how to accomplish basic, standard, everyday functionality in the new layout. It's too inefficient. I leave all that bleeding edge stuff to other distros / graphical environments.
The whole "get used to it" attitude might have been Gnome devs, or GIMP devs, or both, or someone else who decided to fix what wasn't broken. I'm not really clear on which part of the infrastructure introduced the changes or decided everybody needed them. Sorry if it sounded like I was blaming XFCE developers - I hadn't ever thought it was XFCE changing or making such decisions, hence why I had said, "And that XFCE is affected because it has dependency on..."
EDIT: I think I may have mis-interpreted your first comment. The change from CentOS 6 to CentOS 7 was a change from gnome 2 to gnome 3.
Yes indeed. XFCE was beautiful before that point. After that update, there have been a number of annoyances. Gnome System Monitor has lost much of its functionality as its menu vanished. The screenshot program has gone all forum-"Top Posting"-style, with the response buttons appearing first at the top of the window, before you have even read down as far as what those buttons are committing you to. Various other programs have widgets leaking into the titlebar. The window border has become impossibly narrow so that resizing windows is a chore. Scrollbars have lost their arrows to nudge content up/down by a line. Some Save/Open dialogue boxes seem to have been stripped down. It seems extensive.
Reading other users' approaches (such as Linus's) involved overlaying themes or extensions with extra bits and pieces to bit-by-bit restore each aspect of the broken usability. It seems labour intensive, and likely a constantly incomplete solution where the graphical environment will occasionally pop out a, "and another thing you missed!"
So I was hoping simply redirecting XFCE's calls from whatever specific back-end dependency is responsible for propagating all the new GUI features, back to a different version or fork of that dependency, would reinstate all the familiar behaviour in one thorough, consistent, system-wide assignment. Is this overly simplistic?
Xfce is still mostly GTK2 - but the next major release is going to be a port to GTK3.
Is that something that will make it (even) harder to retain normal traditional windowing experience? Because this:
Most of the core Xfce apps are still GTK2. I believe Mate is also based on the same GTK2 libraries. Much of this, however, depends on the application that you use. If the application has been ported to GTK3, then it will use the GTK3 libraries.
suggests to me that each executable uses the libraries it wants to, regardless of what the desktop environment itself as a consistent presentation system is using. If that is so, then my query about getting XFCE to use (and enforce everything else using) GTK2 isn't feasible. I'd need to uninstall GTK3 (if that's the source of the problems) to coerce every individual executable to "fall back" to using GTK2 interfacing.
Would that work or is that overly simplistic? I really don't want to have to roll back to using old versions of programs if I don't have to (I mean, older than the "old" versions RHEL/CentOS has endorsed: EPEL's XFCE is v4.10).
Offline
I agree mostly with what you're saying. A few comments:
1. GTK2 is really a dead project. There isn't any new development being done - only some bug fixes. GTK3 is the successor.
2. The use of GTK2 or GTK3 is dependent on how the application is developed and compiled. If the application is developed using only the GTK3 toolkit, then you have no choice but to use GTK3.
3. GTK3 apps do not fallback to GTK2. If you uninstall GTK3 (which your package manager will probably prevent because of dependencies), then those apps will cease to work.
4. Xfce is moving to GTK3. I don't believe this decision will be reversed. When the next major release is complete, according to the roadmap, it will be built using the GTK3 toolkit (like Gnome 3). It will in all likelihood continue to support and run those apps that are compiled using GTK2, but its main components will be developed in GTK3. And because GTK3 is an actively developed toolkit, it will offer some more flexibility in development. Xfce itself won't change much. It will still have the same desktop metaphor - it will just be built with the more modern GTK3 libraries.
5. I understand that there are two distinct camps when it comes to GTK3/Gnome 3 (hate it or love it), but it is an actively developed toolkit which from a development perspective, is a huge plus. Yes, the Gnome 3 developers are sometimes single-minded and make controversial decisions, but they are committed, driven and focused on a vision - that's their prerogative. GTK3 isn't necessarily a bad thing. As you've mentioned, the Gnome 3 interface can be made to look more Gnome 2 - like using extensions. And also have a look at the desktop paradigm that the Cinnamon DE has come up with using GTK3 libraries.
6. If you do a CentOS 7 minimal install and then install Xfce from the EPEL repositories, you end up with a more GTK2-like interface (mainly because you haven't installed any of the Gnome 3 GTK3 apps). You can also consciously choose GTK2 apps over GTK3 apps (gnome-system-monitor vs xfce4-taskmanger) to prevent the inconsistent user interface issues from interfering.
7. I've never used it, but apparently CentOS 7 has an Xfce 4.12 COPR repository.
Please remember to mark your thread [SOLVED] to make it easier for others to find
--- How To Ask For Help | FAQ | Developer Wiki | Community | Contribute ---
Offline
Xfce itself won't change much. It will still have the same desktop metaphor - it will just be built with the more modern GTK3 libraries.
which means it will change much eg: have a look a parole and you see how many has already changed(menu icons, shortcuts, icons...)
have a look at the gtk roadmap to see what will change in future releases:
https://wiki.gnome.org/Projects/GTK+/Roadmap
* Deprecate grabs, add popup windows with grab semantics
* Notebook replacement with stack <-- so no tabs anymore just osxlike button bars...
* A new combobox <-- again? why?
* In-app notifications
* Deprecate GtkButtonBox
not to mention that gtk3 themes and api will break in every minor release, means: even if Xfce 4.14 needs Gtk 3.14 the apps won't (work/look correctly;print errors) with gtk <>3.14 so there must/should be a new xfce release with every new gtk3 release.
the Gnome 3 developers are sometimes single-minded and make controversial decisions, but they are committed, driven and focused on a vision
I guess the main issue is that GTK is nowadays developed under the GNOME team and their vision is to make GTK the GnomeToolKit, other environments doesnt matter. Anway, almost all important/more complex applications use a different toolkit now for that reason.
And also have a look at the desktop paradigm that the Cinnamon DE has come up with using GTK3 libraries.
Forks like Cinnamon, MATE, budgie <add more here> - are great examples how dumb the decisions of GNOME teams are and how much they drifted away from its user base.
Last edited by sixsixfive (2016-02-25 15:41:16)
Offline
2. The use of GTK2 or GTK3 is dependent on how the application is developed and compiled. If the application is developed using only the GTK3 toolkit, then you have no choice but to use GTK3.
Ahh I see. OK. So each program pretty much ignores what precedent the desktop environment running has set for the user experience, and just blitzes ahead and draws its windows its way - based on whether GTK2 / GTK2 / QT was linked.
That explains the inconsistent / heterogeneous pattern I'm seeing. Thanks.
It also answers my question about whether it's as simple as the DE prescribing a different drawing library for all X programs launched under its auspice. I see now that's outside its prerogative.
5. I understand that there are two distinct camps when it comes to GTK3/Gnome 3 (hate it or love it), but it is an actively developed toolkit which from a development perspective, is a huge plus.
I guess the main issue is that GTK is nowadays developed under the GNOME team and their vision is to make GTK the GnomeToolKit, other environments doesnt matter.
Sure. That people are releasing this software means they believe in it, so I was open to trying out whether it could do all the normal things better (more efficiently) than the established status quo. However, my experience unfortunately is that many of them it can't do at all (let alone better) ... and that's by design.
I think innovation is a great thing, and it wasn't long ago that famous skeptics were poo-pooing the WIMP paradigm but it has taken off. So sometimes a group with a vision need to make bold steps away from convention and reimagine "the way things should be". 9 times out of 10 that falls flat but rarely it hits jackpot.
If the goal is to make people say, "hey that looks weird and totally different from everything else I've seen before! I am so curious I want to devote hours / days / weeks to figuring it all out!" then enthusiasts (and newbies with it as their first experience) will love it and wonder what all the fuss is about. For anyone else though, every step of alienation raises the barrier to entry by obstructing tried and tested routines.
I'm sure they've done their cost-benefit analysis. It's a big deal to discard what works in favour of what you imagine might be better under certain conditions, so they must believe in their dream. And passion is what pays for free software after all!
6. If you do a CentOS 7 minimal install and then install Xfce from the EPEL repositories, you end up with a more GTK2-like interface
Good plan. I suppose I could also tell yum to explicitly exclude GTK3 from any installations, thereby red-flagging anything I naively try to install with a GTK3 dependency (as yum will fail to install it due to missing dependencies).
Forks like Cinnamon, MATE, budgie <add more here> - are great examples how dumb the decisions of GNOME teams are and how much they drifted away from its user base.
Maybe I've misunderstood but it sounds more to me like they have forked the userbase. They have struck out on their own in pursuit of their hopes and dreams, and invited anyone who wants to come along for the ride to join. If people want to row their own boats that's great, but the gnomes are rowing the Gimp3 boat.
I like it that there is diversity to choose from and maybe one day good things will come from the new direction. I personally find the loss of functionality crippling while it's still so incomplete. I find it reassuring when the computer acts predictably and supports my workflow. I'll set up a minimal + EPEL + XFCE on a spare drive and see how effective at preserving normalcy that is.
Xfce 4.12 COPR repository
I am somewhat wary of repos not endorsed by CentOS. COPR might be fine, trustworthy folks - I just don't know, and don't know how to find out reliably. XFCE 4.10 should continue to be OK - now you have helped me understand what exactly was causing all the glitches. Much obliged!
Offline
> I think innovation is a great thing,
its not innovation since except the shell everything else is copied from OSX
>So sometimes a group with a vision need to make bold steps away from convention and reimagine "the way things should be".
But that so called vision should have no influence on other projects. Changing the (to that date)defacto default toolkit to support only one desktop is not a vision its a tactic i usually see from MS or Apple - but I didn't expect anything else from a former GNU project that starts talking about trademarks...
Its also weird thet the GNOME project never cared about their user wishes but when debian and red hat announced that they would not ship GNOME as their default desktop environment changes where possible.
>I like it that there is diversity to choose from and maybe one day good things will come from the new direction. I personally find the loss of functionality crippling while it's still so incomplete. I find it reassuring when the computer acts predictably and supports my workflow. I'll set up a minimal + EPEL + XFCE on a spare drive and see how effective at preserving normalcy that is.
forks have nothing to do with diversity, they are a bitter split. Its usually done when the main project don't listen to its core users(eg: by rejecting feature requests). currently we have ~3.5 GNOME forks that try to archive the same goal, currently the only archived goal is that Qt is becomming more and more popular...
Last edited by sixsixfive (2016-02-26 20:18:11)
Offline
Pages: 1
[ Generated in 0.010 seconds, 7 queries executed - Memory usage: 610.57 KiB (Peak: 627.41 KiB) ]