Linux Mint Developer Forks Gnome 3

Long-Term? by headkase · 2011-12-21 11:22 · Score: 4, Insightful

How long can he keep it up and what about long-term compatibility with GNOME 3 apps? Eventually I'm sure their "lineage" will drift far enough apart that you're either pulling in multiple families of libraries that do the same thing or you get GNOME 4 apps that don't work on Cinnamon 4 and vice-versa.

Anyway, I'm typing this on Arch Linux 64-bit with GNOME 3.2.1 and a few (needed!) shell extensions. I find it fine and I thought I would be a GNOME 3 hater but I'm actually not.

--
Shh.

Re:Long-Term? by erroneus · 2011-12-21 11:46 · Score: 5, Insightful

If no one tries, it never happens. GNOME3 with its GNOME shell sucks ass and turns a desktop into a netbook candy toy interface. Perhaps if they are successful at giving it older/better functionality, then after the public appreciates it, they might merge it in with GNOME3 in some way.
GNOME3 shell extensions need to be better managed and maintained. My second attempt at using Fedora... this time 16 is still a failure as far as I'm concerned. The extensions idea is nice but it doesn't inherently manage the options. What resulted was a GNOME3 shell that wouldn't load unless I kept shuffling extensions to try to get what I want. GNOME3 and its extensions interface does not account for or manage the extensions which are present and running. (It seems kind of obvious to me that when a UI element is being manipulated in some way by an extension, a 'lock' preventing other extensions from acting on it should be created and enforced.)
I have heard there is now some sort of central extensions repository and I hope it alleviates the extensions mess I experienced but I think over this holiday time, I am going to load CentOS 6.x instead of Fedora.
Lately it seems software projects are refusing to listen to their users and it shows. GNOME3's shell, Firefox and SME server are three that have affected me in a large way and none of them seem interested in listening to the feedback.

Keep away the UI "designers"! by Anonymous Coward · 2011-12-21 11:27 · Score: 3, Insightful

Whatever they do, they need to make sure that they do everything in their power to keep away the self-labeled "UI designers" who have fucked over GNOME, Firefox, and numerous other major open source projects lately.

These people may think they know how to create a usable UI, but experience shows that they have no fucking idea what they're doing. Just look at how damn unusable Firefox is these days. The menus are gone, the status bar is gone, the protocol in the URL bar is gone. It's hard to get anything done in Firefox. Sure, I can dig through the settings to re-enable those things that should never have been disabled by default, but that takes far too much effort. It's easier to ditch Firefox. The same goes for GNOME. The "designers" fucked up its UI, and now it's unusable. Now we see real software developers trying desperately to fix the situation.

It's more harmful to an open source project to let them contribute than it is to constantly shut them down. Do not respond to them on mailing lists or IRC. Do not let them get any sort of commit rights. Close any "usability" bugs they open. Do not let them participate in any way.

Only let actual software developers create UIs. They may not be pretty, but at least they'll be functional and much better than anything "designed" by the "UI designers" that have ruined GNOME and Firefox.

Re:Keep away the UI "designers"! by 0123456 · 2011-12-21 12:05 · Score: 5, Insightful

Or perhaps you're talking about the status bar. Again, something I can't believe anyone would notice or care about. A largely blank, useless bar that was practically only good for previewing link URLS was removed from the GUI and replaced with something smarter. Again, how is this a major change?
Useless to you, perhaps. But the replacement is a kludge for tiny screens that's a horrible mess on a desktop with a decently sized monitor. I find it's contnually covering up things I want to click on all for the sake of not 'wasting' a few pixels on a 1920x1080 monitor; it's annoying, it's ugly and it provides no benefit over the old status bar.

You make it difficult for people who actually have good, valid criticism and feedback of GNOME 3, etc., to be heard, because you dilute the discussion with completely bizarre, emotional, thoughtless statements.
Or perhaps you just don't bother to understand why people want these 'useless' features.
Re:Keep away the UI "designers"! by fnj · 2011-12-21 12:16 · Score: 2, Insightful

The more minor part of the hostility is that the defaults are stupid. Sure, you can change them if you want, but why should the 99% of sane people have to go to even that much trouble just to cater for the 1% of idiots who like the tabs in the wrong place, and the utterly pointless Amazing Invisible Menu?
The destruction of the status bar was just plain stupid, and there is no option to bring it back. You have to install an addon to regain elementary usability because of this moronic decision.
The major part of the hostility is because this is all a sign that all the developers have now become too superior (in their own minds only) to adhere to a common user interface, signaling a trend back to the wasteland of every app being completely idiosyncratic - a trend that is completely destructive to usability and learning/training.
Re:Keep away the UI "designers"! by vashfish · 2011-12-21 19:45 · Score: 1, Insightful

All you've done is list the differences and assert they are dumb changes. You have said nothing to actually show that these differences are detrimental to usability...
Re:Keep away the UI "designers"! by aaaaaaargh! · 2011-12-22 02:20 · Score: 3, Insightful

I'd have no problem with self-proclaimed UI designers as long as they'd respect the following very basic "rules of thumb":
* Every command can have a keyboard shortcut.
* Issuing a command immediately provides visual feedback (always and with absolutely no visible delay, even menu items should blink).
* While a command is issued or visual feedback is given other commands can be issued without delay, provided that processing has not become very slow and the queue becomes long (the latter must be avoided at all costs by using suitable programming techniques and data structures but of course sometimes a machine is just doing too much work).
* Important commands are no more than one mouse click away, less important ones 2 or a maximum 3. There is really no need for an UI where you need to click or open 3 different menus/views/buttons/windows to get anywhere.
* All visible GUI elements such as toolbars, panels, buttons are freely configurable both in their content and their spacing and place.
* All interface elements can be selected and used with the keyboard or there are equivalent keyboard commands.
* Windows and interface elements always remember their settings such as position, size, etc.
* Modal dialogs are avoided as much as possible.
* Instant/live update of the results of search fields is welcome, but then it must be instant--no delay.
Voila! A working GUI...at least in my opinion.

Re:You're... by muszek · 2011-12-21 11:28 · Score: 5, Insightful

A year or two ago everybody was happy with Gnome. Just Gnome, we didn't have to call it Gnome 2.x. Now we have Gnome 2.x, plain Gnome 3.x, Unity, Mint Gnome Shell Extensions, MATE and now another kid on the block... what the hell went wrong?

I'm still happily using Gnome 2.x (on LMDE), but it won't last forever :/

Re:And I care because ..? by pseudofrog · 2011-12-21 11:41 · Score: 5, Insightful

So you're like the "I don't own a TV" guy.

Nobody cares that you don't care. Get over yourself. Seriously.

Awesome. by Improv · 2011-12-21 12:06 · Score: 5, Insightful

The components of GNOME3 are mostly great, but the overall experience is terrible; the thing feels like it's designed for tablets, or as part of a blue-sky interface experiment. They took out most of the options that would've let people make it usable again, and have showed hostility to existing apps and user priorities (screensavers are so 90s? Really?). Compatibility with apps written against GNOME3 libraries is great, especially if we can get most of the good stuff from GNOME2 back.

If the GNOME Foundation doesn't want to deal with this, they should get rid of a lot of the people who made the poor decisions that led them to release a terrible, constraining product.

--
For every problem, there is at least one solution that is simple, neat, and wrong.

Re:GNOME has always been fucked up. by Anonymous Coward · 2011-12-21 12:11 · Score: 0, Insightful

I haven't checked out GObject, but if it's anything like GtkObject then you are just badly informed. The code base is immaculate in anyway you look at it, top-notch formatting and commenting, excellent use of patterns, easy to make sense of. Arguably one of the best pieces of open source code out there, judging by code quality.

Re:You're... by somersault · 2011-12-21 12:16 · Score: 4, Insightful

What? This fork sounds like it is entirely to create the best end product. Keep the parts that have improved in Gnome 3, but get back the good stuff from Gnome 2.

I've been using Mint for months, it's a good OS. It seems to me that this guy has his head screwed on right - as opposed to those who are desperate to turn their desktop OS into something that only makes sense on a tablet.

--
which is totally what she said

Re:You're... by smash · 2011-12-21 12:56 · Score: 5, Insightful

Linux/Unix desktop environments at the moment appear to be all about the colour of the bicycled shed, rather than things that ACTUALLY matter to end users / developers such as a stable ABI. Example: in Windows i can run most applications all the way back to the mid 90s without major problems. OS X has even carried compatibility with old apps for at least 5 years, and its been through a major operating system redesign and CPU architecture shift.

Can I do that with the free unix desktop? Sure, vanilla X apps probably work, but every major rev of KDE (haven't tried old gnome apps on newer gnome versions) breaks heaps of old apps. Every version of KDE or Gnome i have ever used since both projects began (i remember compiling KDE 1.0 and QT from source and being impressed :)), i have found "wierd" shit where i can make part of the UI crash or errors thrown on screen.

Please: stop fucking around with eye candy and the colour of the bicycle shed. Debug what you have, get it stable and THEN go about adding new stuff. Just because Windows or OS X has new feature of the month, it doesn't mean you need to kludge a clone of it on top of your DE within 2 weeks in some shitty half-assed way.

"Usability" of a UI is to a certain extent, bullshit. Most users can adapt to design decisions made on your environment. Apple knows this - yes, I wish i could customise the OS X desktop a bit more, but at the end of the day the fact that I can't is no major deal-breaker. Because it actually works. Yes, UI testing can make soemthing a little nicer to use - but if it is full of bugs, crashes, breaks your old apps that you like and generally misbehaves, then all that usability testing and research is WASTED.

I didn't mean this to turn into a big unix-desktop rant, but i've been really wanting to like the unix desktop since 1995. Some aspects of it, I do love. But since the days of say, KDE2 (or gnome equivalent - essentially when we got a usable file manager style desktop), there's been very little actual progress in real world usablity that I can see. Sure, there's new eye candy. Whoopie. Can it help me get shit done better? Not really.... progress appears to have stagnated.

--
I run: Windows, OS X, Linux, FreeBSD. Just because you have a hammer, doesn't mean everything is a nail.

Agreeing with every point here, except one... by Anonymous Coward · 2011-12-21 13:01 · Score: 5, Insightful

Modularity is a good thing. It's not cutting up things into a lot of small modules (aka "libraries") that's the problem.
It's doing it wrong.

Look at the typical bash shell and GNU utilities we all use every day. They are hundreds of small executables, libraries, etc. But they are not a mess. They all do one thing, and do it well. That's part of the UNIX philosophy, and for a reason.

And about KDE: A monoculture is never good. Even two are not enough for a healthy ecosystem. And what's the problem with forking anyway? It doesn't hurt anyone,and nearly has no overhead. (If you use git and know how to use it.) The fear over "fragmentation" is entirely delusional and pointless. We are not one of those idiotic "everybody must follow his party line, no matter what" systems. We are not a US two party system.
In fact, I think every user should have his own fork by default. Where "fork" can mean anything from an empty patch set to fundamental major changes. And everybody should just be able to "subscribe" to whoever else's personal fork, implicitly making that someone else a "distributor" without having to do anything special. So that natural leader/follower structures can arise, and nobody can force anything on anyone.
(Sorry for sounding so angry. I don't mean to say this in a attacking way. I'm just a bit beside myself right now for completely unrelated reasons, and can't switch it off. Your post is still 95% in harmony with my opinions. :)

Also, there is one additional thing you missed: The moment "desktop environments" for Linux started to forget the UNIX philosophies, abandoned the concept of "everything is a file", and chased the Windows and OS X, they were full of FAIL and lost anyway. (There's no file system for your GUI, is there? You can't cat /proc/pid-6939/window-2/grid-3-2/textarea-2. It's all monolithic Windows-like "applications". You can't use a GIMP brush in OpenOffice, you can't use the same text layouting engine for OpenOffice, Firefox and GIMP, etc, etc, etc. It's all just deeply deeply anti-UNIX, harming code re-use, customizability, modularity, and most of all usage efficiency. And all for the sake of Joe Sixpack, who is a retarded dick anyway, please please loving you... but not really loving you, since you deformed yourself until you talked like a Windows/OSX and walked like a Windows/OSX, and he really only loves you when you have become more Windows/OSX and Windows/OSX itself. In other words: He still won't love you. So quit lying and be yourself! Same as the typical problem geeks have with women, interestingly.)

Re:Agreeing with every point here, except one... by lennier · 2011-12-21 15:22 · Score: 5, Insightful

The moment "desktop environments" for Linux started to forget the UNIX philosophies, abandoned the concept of "everything is a file", and chased the Windows and OS X, they were full of FAIL and lost anyway. (There's no file system for your GUI, is there? You can't cat /proc/pid-6939/window-2/grid-3-2/textarea-2. It's all monolithic Windows-like "applications". You can't use a GIMP brush in OpenOffice, you can't use the same text layouting engine for OpenOffice, Firefox and GIMP, etc, etc, etc. It's all just deeply deeply anti-UNIX, harming code re-use, customizability, modularity, and most of all usage efficiency./p>
YES. THIS. EXACTLY THIS.
I really, really want a Unix desktop which actually implements the Unix philosophy, and very much want my windowspace to be exposed as a file system (or rather, as a VERY loosely / completely untyped object system. And no, sadly JSON objects don't quite cut it, which is a problem since we're baking JSON into the Web. Lua objects would probably work though; they're pretty nice, and it interoperates well with C.)
I want the ability to, just as you say, reuse objects and components from one "application" inside another. In fact, I want to completely erase the concept of "application"; I just want a robust store of data, as a set of fine-grained untyped objects/collections, and then various views or functions over that data. And then yes, publish any part of my data/function/object hierarchy in a safe, standard way to a net-wide repository as a sort of mini-distribution, and safely import subsets of other people's stuff into mine.
See, there's a whole lot of nonsense busywork we're currently doing in the system administration space which duplicates and triplicates stuff we've almost solved in the programming space - only badly, and without interoperability. For example, what is a zipfile but an untyped object containing other files? What's a directory but an almost-but-not-quite-the-same object? What's a filesystem but again, an almost-but-not-quite-the-same thing? What's a version control "commit" but the same thing as an RPM/DEB patch, except implemented differently? What's a "distribution" but something that ought to just be an RPM of RPMs? And what are SQL "databases" and "tables" but again, objects containing sets of data, and why do I need multiple different incompatible formats for each one?
So we have, at the OS/system level, these various different implementations of the idea of "structured object", but not really done sensibly; for one thing, there's this very archaic concept of a single shared filesystem which is very much like the old pre-1960s (FORTRAN and COBOL) programming concept of global variables. In programming languages, we moved past global variables toward structured sets of local variables when C came along; but we didn't at the filesystem level. This leads inevitably to easy corruption of a system: run one installer with root priviledges, and it has access to your entire root namespace on your hard drive. Our systems shouldn't really, in 2011, be structured in such an old-fashioned way.
So at the OS layer we have "files and directories as objects". Then we have a process-management layer over the top: libraries, processes, threads. Then we reimplement the idea of "object" AGAIN (but in a non-interoperable way) as various "software component" frameworks (which of course install into a global per-system namespace, stomping all over each other to create DLL hell): COM objects, XPCOM objects (Mozilla), UNO objects (OpenOffice), CORBA objects, COM objects, KParts objects, GObject objects, .NET assemblies, Java namespaces. And a "registry", which reimplements the filesystem, to handle invocation of the software componentry. Then we reimplement the idea of "object" a fifth time as non-persistent programming-language objects which only exist in a single running process space : C++ objects, PHP objects, JSON/Javascript objects, .NET objects, Java

--
You are not a brain: http://books.google.com/books?id=2oV61CeDx-YC
Re:Agreeing with every point here, except one... by rev0lt · 2011-12-21 17:53 · Score: 3, Insightful

The philosophy of "everything is a file" is a naive one. It worked well in the 70's, where you either had text files or binary files (and a folder is a special binary file), and most storage units didn't have more than 10 000 files. Today, you have multiple different kind of containers with multiple types of information. As an example, think of a video file. Should the metadata properties also be accessed as a file? Should the sound and video be accessed as different files/streams? And how about when both streams are interleaved? And the keyframe index, should it be accessed as a file also? Should JPEG extensions (such as thumbnailing) be scanned and exposed as a file? And how about metadata referring to non-available applications, such as Photoshop Exif entries? And even if everything was a file, how would that help you to find that 300x700 portrait you have of your mom, taken somewhere last year?
We are moving away from container-based storage units to metadata-based storage, precisely because the notion that everything is a file is quite limited. And these limitations aren't even new - symbolic links are in some ways a hack that breaks that base approach - you can refer to the same object from multiple different container, which - by itelf, is a rudimentary relation mechanism. I won't even mention ACLs - you access a file, but the system actually opens (at least) 2 files in many implementations, because the "file" notion doesn't comprehend accountability or complex ownership.
The big players (Apple and Microsoft) have been moving away from file-based storage for years, and on to metadata-based stored approach. And no, afaik this isn't something you can easily slap over an existing filesystem.

Also, the same concept you praise is contrary to the integration you preach - each vendor should implement the funcionality they need over the archaic "file" concept, as there is no "one size fits all" when it becomes to content decoding, and for the base libraries to actually be useful, they would have to be generic (think of the file api right now).
We have huge bloated frameworks because different people has different needs, and processing power is cheap - cheaper than development time. That's what having a programmable device is all about - being able to write your own bloat how you think it should be implemented, instead of eating the other person's bloat.
Re:Agreeing with every point here, except one... by DrXym · 2011-12-21 22:53 · Score: 3, Insightful

I think the thing Unix got right was to ship a bunch of tiny tools that could be strung together via pipes and redirections to construct something useful. It's a mind bogglingly useful approach, and the tools are powerful especially compared to the half assed tools that plague DOS derivatives to this day.
That said there is no denying how "organic" some tools are. There is no consistent syntax between tools, and some tools are arcane or implement arcane default settings. I also have a love / hate relationship between bash, gawk and perl and constantly have to relearn these bastards when I need to write a script because they're almost write-only languages and virtually unmaintainable once they grow beyond a certain size. I once had to port a 5000 line cgi perl script which could generate 6 disparate web pages into Java. It took six months to unpick and reimplement.

Re:You're... by s4m7 · 2011-12-21 13:40 · Score: 5, Insightful

Linux as a whole (kernels, UIs...) has turned into a developers dick size contest. Everybody wags their own, nobody debug/documents/supports appropriately for end users.

Linux as a whole, is the kernel. The kernel. There are different versions, patches, etc. but it's one kernel.

Maybe you mean open source as a whole?

Maybe you mean software as a whole. That would make a whole lot more sense. Except it hasn't "turned in" to anything... it's always been that way.

--
This comment is fully compliant with RFC 527.

Why doesn't Gnome get it? by walterbyrd · 2011-12-21 14:23 · Score: 4, Insightful

The community has beating Gnome over the head for months now. But Gnome stubbornly refuses to go back to their less FUBAR interface.

What the hell is wrong with them?

Oh well, at least there's forking.

Re:Why doesn't Gnome get it? by lennier · 2011-12-21 15:46 · Score: 4, Insightful

Oh well, at least there's forking.
Forking is the answer to borking.

--
You are not a brain: http://books.google.com/books?id=2oV61CeDx-YC

Re:You're... by JDG1980 · 2011-12-21 16:32 · Score: 4, Insightful

I for one don't understand why people get all emotionally attached to their old UI.

Muscle memory.

Once you've gotten used to using a specific UI for years on end, the commands are basically hard-coded into your body. Changing this takes a lot of time and effort and you will often find yourself automatically doing things the old way.

Re:GNOME has always been fucked up. by icebike · 2011-12-21 16:47 · Score: 4, Insightful

I can't disagree with your take on the politics. I do take issue with the technology.

Very odd. After making that statement you go on to validate just about everything the GP said.
You get modded up for starting an argument, but before you've written 2 paragraphs you've agreed with the other guy by just using different words.

Are you guys brothers?
My brothers used to fight against each other on but the same side of the argument a lot too.
Seems like the arguments always ended with "Ok then". To which the other replied "Fine".

--
Sig Battery depleted. Reverting to safe mode.

Re:GNOME has always been fucked up. by bonch · 2011-12-21 17:31 · Score: 5, Insightful

I don't understand the abhorrent removal of features and configurability.

The world has moved on from the 1990s desktop. Windows 8 is replacing the Start menu with a radical new touch interface, and OS X Lion has adopted iOS features. Mobile operating systems that look and feel nothing like desktop operating systems are a huge hit. The 1990s desktop paradigm is dead in the mainstream, and Linux software developers are trying to progress computing forward by appealing to people outside of tech forums. That means reducing the insane amount of configurability and feature-itis that often ails Linux desktop software.

You may consider it abhorrent, but you are a minority. The rest of the world uses computers simply as a tool, not as a hobby.

Re:GNOME has always been fucked up. by Lisias · 2011-12-21 18:23 · Score: 4, Insightful

PLEASE MOD PARENT UP. Abusive moderation detected.

I do not agree with what it's saying, but the opinion is valid and the guy didn't do any offense!

--
Lisias@Earth.SolarSystem.OrionArm.MilkyWay.Local.Virgo.Universe.org

Re:GNOME has always been fucked up. by Pi1grim · 2011-12-21 20:01 · Score: 5, Insightful

I use my computer as a tool. And as such it must be efficient. If make a hammer from soft rubber justifying that "common user might get hurt when using metal hammers" then you are making a goddamn toy, not a tool. And that is exactly what Gnome crew is doing. I get that they are trying to make an impressive interface for tablet PC, but why should regular PC users suffer? Would it hurt them to go play in a "Gnome-Tablet" version of the interface without crippling the desktop version? Those that would not need customizability and would be happy with all the bling could play with the new interface all they want and those that do any actual work could continue to do so without switching DEs.

Re:GNOME has always been fucked up. by xtracto · 2011-12-21 21:19 · Score: 3, Insightful

QT was amazingly good for C++, Gnome couldn't compete

But the idea was to make an *open source* desktop environment. I am sure a lot of C++ programmers would have gone helped an effort to make an open-source version of QT.

--
Ubuntu is an African word meaning 'I can't configure Debian'

Re:GNOME has always been fucked up. by kangsterizer · 2011-12-21 23:47 · Score: 5, Insightful

Yes and no.
It's the "Silicon Valley effect".

That is, some guys change stuff in the UI and declare it's where the innovation is. It's not that it's good. It's just different, and backed with strong advertising (Apple, etc).

Then others feel they *have* to follow a similar path to survive and "have something new to announce". Otherwise, you're not news worthy, etc, etc.

I also call it being blind :P

It's fine to copy concepts, but copy the ones that are actually good.

Re:GNOME has always been fucked up. by Rich0 · 2011-12-22 00:17 · Score: 5, Insightful

The rest of the world uses computers simply as a tool, not as a hobby.

The rest of the world uses windows or OS X.

I find it odd that FOSS advocates are willing to reject much of what they find useful with desktop enviornments to make them more appealing to people who will likely never use them...

Slashdot Mirror

Linux Mint Developer Forks Gnome 3

28 of 314 comments (clear)