| |||||||||||||
Otte: staring into the abyssOtte: staring into the abyssPosted Jul 27, 2012 18:34 UTC (Fri) by drag (subscriber, #31333)In reply to: Otte: staring into the abyss by DiegoCG Parent article: Otte: staring into the abyss
The problem is there is no path forward.
Until Linux distributions finally break themselves and are willing to make a changes necessary to be a good desktop OS no Linux DE is ever going to progress beyond were it is at now.
It's not healthy for a project to be in a stale environment with no new blood or new interest. There really isn't anything Gnome can do about it right at this moment.
(Log in to post comments)
Otte: staring into the abyss Posted Jul 27, 2012 20:06 UTC (Fri) by khim (subscriber, #9252) [Link] Until Linux distributions finally break themselves and are willing to make a changes necessary to be a good desktop OS no Linux DE is ever going to progress beyond were it is at now. This is not true at all. We have not one, but two viable contenders. Sure, they are not yet there. Even Gnome3 is better… for now. But they are improving and, more importantly, they offer viable desktop platform for the third-party developers. And in the end third-party developers make or break the desktop platform.
Otte: staring into the abyss Posted Jul 27, 2012 22:27 UTC (Fri) by BradReed (subscriber, #5917) [Link]
It is strange to me that everyone seems to say Gnome and KDE are the only DE choices. Why are things like XFCE or e17 dismissed?
Otte: staring into the abyss Posted Jul 27, 2012 23:09 UTC (Fri) by drag (subscriber, #31333) [Link]
Because e17 took several eons to accomplish anything and xfce is mostly gnome with pieces removed and a handful of applications.
Otte: staring into the abyss Posted Jul 28, 2012 19:57 UTC (Sat) by cmccabe (guest, #60281) [Link]
Xfce is not "GNOME with pieces removed." It's a completely separate code base from GNOME. You can check out the code at http://git.xfce.org/. Xfce does depend on GTK+, but that's just a graphics toolkit.
If anything, GNOME and Xfce are farther apart than they've ever been. Just compare a default install of GNOME and a default install of Xfce, and tell me how many similarities you see. Not many.
I view Xfce as the working man's desktop because it enables you to get stuff done with a minimum of yak-shaving. Rather than focusing on "the semantic desktop" (like KDE4) or chasing dreams of Linux tablets (like GNOME3), Xfce focused on delivering a usable desktop with no drama. And it has succeeded. It may not be as exciting as a catastrophic rewrite-from-scratch or as trendy as HTML5, but it's actually usable. I hope more distros consider using it as the default desktop.
Another thing Xfce gets right is not re-inventing wheels. GNOME has its own office suite (GnomeOffice), its own IDE (Anjuta), its own CD ripper (Sound Juicer), and on and on. Xfce takes the approach that if a perfectly usable program already exists for a task, it doesn't need to be re-invented.
P.S. I've done the "old-school WM" thing before too. I used xfvwm for a few years, and even used twm. None of those environments was as productive as Xfce because there was inevitably a yak to be shaved-- a configuration file that had to be edited by hand, or a bash script that had to be tweaked. Xfce is a great project and it's really worth a look, if you haven't tried it.
XFCE Posted Jul 29, 2012 14:52 UTC (Sun) by dskoll (subscriber, #1630) [Link] +1 We use Debian Squeeze at work. When we move to Wheezy, I'll be transitioning all the non-technical users from GNOME 2 to XFCE because I think that transition will be easier on them than GNOME 2 to GNOME 3.
XFCE Posted Jul 29, 2012 16:18 UTC (Sun) by patrick_g (subscriber, #44470) [Link]
Early 2013, a lot of Wheezy users will dislike Gnome 3 and will switch to Xfce. It's because of that it's so sad to have only the 2 years old Xfce 4.8 in Wheezy instead of the one year old Xfce 4.10.
XFCE Posted Jul 29, 2012 19:09 UTC (Sun) by dskoll (subscriber, #1630) [Link] I'll take old XFCE 4.8 over new GNOME 3 any day. Besides, as far as the end-user experience goes, I don't think there's that much difference between XFCE 4.10 and XFCE 4.8.
XFCE Posted Jul 30, 2012 22:00 UTC (Mon) by mbiebl (subscriber, #41876) [Link]
It might be easier for you to just use GNOME Classic (fallback mode)
XFCE Posted Jul 31, 2012 0:50 UTC (Tue) by bronson (subscriber, #4806) [Link]
There have been mixed messages about the future of fallback mode... It might not be the best choice for stable, long term deploys.
XFCE Posted Jul 31, 2012 2:36 UTC (Tue) by dskoll (subscriber, #1630) [Link] Not. GNOME classic means I have to go do configuration on all the boxes, whereas switching to XFCE is simply apt-get install xfce4.
XFCE Posted Jul 31, 2012 19:09 UTC (Tue) by mbiebl (subscriber, #41876) [Link]
This doesn't make sense.
Since you are apparently using Debian, just select the "GNOME Classic" session in your display manager and you're done. No configuration required.
XFCE Posted Aug 1, 2012 17:09 UTC (Wed) by dskoll (subscriber, #1630) [Link] Hmm, OK. Didn't know that. However, I think XFCE has more of a future than GNOME Classic and I don't want to have to force my users to change again in a year or two's time.
Otte: staring into the abyss Posted Jul 31, 2012 11:45 UTC (Tue) by rossburton (subscriber, #7254) [Link]
"its own CD ripper (Sound Juicer)"
I wrote SJ originally because there wasn't anything else. At the time the cutting edge GNOME-friendly ripper was Grip, and after wasting a week of constant ripping because I typo'd one of the patterns I decided that enough was enough and wrote a replacement.
SJ is part of the GNOME release but it hasn't actually depended on anything GNOMEy for some time now, so the current releases will fit in fine once Xfce moves to GTK+ 3 (or use SJ 2.32, and bring in a small library from gnome-media).
I was curious to see what Xfce recommended as a CD ripper. The wiki page recommends an app that appears to be unmaintained now, and the xubuntu metapackage doesn't appear to pull in a CD ripper (maybe I missed it, or it's integrated into something else).
Otte: staring into the abyss Posted Aug 5, 2012 6:52 UTC (Sun) by dirtyepic (subscriber, #30178) [Link]
Is that xfburn? I use it from time to time and it usually does the job.
Otte: staring into the abyss Posted Aug 16, 2012 23:17 UTC (Thu) by cmccabe (guest, #60281) [Link]
Thanks for the information! I'll give Sound Juicer a try next time I have to burn a CD.
Otte: staring into the abyss Posted Aug 16, 2012 23:18 UTC (Thu) by cmccabe (guest, #60281) [Link]
Er, that should read 'rip a CD.' Opposite operation :)
Otte: staring into the abyss Posted Aug 19, 2012 12:56 UTC (Sun) by philomath (guest, #84172) [Link]
That's my experience too. Xfce is the only DE workable for me.
Otte: staring into the abyss Posted Jul 29, 2012 8:19 UTC (Sun) by bluebugs (subscriber, #71022) [Link]
Dismissing E17, because it takes time to do think technically right is quite abusive. First of, E17 first needed a modern cleanly designed toolkit. So a massive amount of work was moved into it, to get the lightest, and certainly the most efficient toolkit. With that toolkit, it was possible to write a new terminal in less than one month : http://www.enlightenment.org/p.php?p=about/terminology . Terminal that is faster and lighter than most existing alternative, but still highly themable.
The main reason E17 was never released is that it reached the good enough state and nobody cared to make the last step. In fact most of our community is focusing on the library instead of releasing application. Thanks now, some people are taking care of that last move. But the good enough and the fact that we were fine in our sandbox is the main reason of this slowness.
So don't dismiss technology if you don't know there story and what they can do.
Otte: staring into the abyss Posted Jul 31, 2012 11:10 UTC (Tue) by ebassi (subscriber, #54855) [Link] First of, E17 first needed a modern cleanly designed toolkit. which is... what, precisely? which one of the bajillion libraries (with runtime loadable modules as separate shared objects), all of them thread-unsafe, with no documentation, and no comments (except for the sporadic 'printf("WAAAAAAA")' ifdeffed out), with a runtime type system based on magic, hardcoded strings and public structs of virtual functions, is the "modern cleanly designed toolkit" you're blabbing about? because I haven't seen anything like that in the Subversion repository (and, really? Subversion?) everything inside the collection of random pieces of junk that is known as "EFL" screams bad design, bad implementation, bad software engineering and release management practices. it's a disgrace that this utter, utter joke of a project has been allowed to go this far, and con lots of money, time, and effort out of users, developers, and companies. have you ever asked yourself why there isn't any documentation on how to create a complex EFL-based application? or, for that matter, why there is no complex EFL-based application at all, except Enlightenment, and now closed sources apps from Samsung and their phones? have you ever wondered why the EFL backend for WebKit is using GLib, libsoup, cairo, gstreamer, and other GNOME technologies? hint: it's because they work, exist, and are designed, developed, and released by competent developers. Enlightenment is a joke, and a bad one at that. the fact that you're using it is utterly inconsequential to its quality, or its development practices. unless you are one of the contributors, to which I would reply that you are a terrible, terrible person, and are directly responsible to two awful weeks of my life in which I tried to work on that horrible code base for my former job, and that I would love to get them back.
Otte: staring into the abyss Posted Jul 31, 2012 14:44 UTC (Tue) by bluebugs (subscriber, #71022) [Link]
Wondering here how you did spend your two weeks, certainly not looking at the right place. Nice troll about Subversion by the way, but that one I will let it to someone else. Let's start just by fact.
Documentation : http://web.enlightenment.org/p.php?p=docs&l=en . As you can see, updated daily and available online easily.
"Bad implementation, bad design, bad software engineering", man you are really shooting yourself in the foot here. Enlightenment is faster, lighter and yet more powerful than any thing you can compare it to. Writing a terminal emulator from scratch took 1 month as a pet project, and beat hands down every other terminal emulator in speed and effect (memory when comparing it with tabless emulator). If that's what you get by bad implementation, bad design and bad software engineering, every one should start doing bad stuff that would help !
As for big application, the main reason is that nobody in the Enlightenment community were interested by replacing existing application. We were having fun in our sandbox, didn't bother doing marketing and didn't try to get more users/developers. You may have notice that we do have around 30 people working permanently on all EFL libraries (when GTK only get one), that's because people in our communities have fun with improving and tunning this libraries. But things are evolving. That's why we now have release and are starting communicating. That's why we do start writing more application. Because we also thing that it's a time that require us to speak. Let see if in two years we don't get big applications written in EFL.
About WebKit, well it was a GTK port that was moved to EFL (That does prove my point to a previous thread where moving to EFL is doable), so obviously not all technology get replaced. And what GNOME technologies are you speaking about ? Cairo, Gstreamer ? Are you sure they where created by and for GNOME ? Are you sure you can be so proud of their current state ?
As for the fact, we did have a Cairo backend for Evas, we scratched it because it was to slow and never was close to the performance we had with our own software backend. The only reason we need it is for vector graphics, and we are close to have an alternative that is already way faster. And Gstreamer, on any PC, VideoLan give better result, it's more stable and use less resource. We do use Gstreamer, because that's the only choice we have in the embedded world today, but frankly on a desktop go with VideoLan !
By reading your comment, I guess you were forced to look at EFL and took it as a personal attack on your own work. Did your management ask you to dumb all your past work because it didn't perform correctly ? Seriously you go on personal attack here, and it really look to me like EFL did beat the code you did write in the past without giving you a little chance to catch-up. I feel sorry for you, but if one day you calm down and really want to learn about how a modern toolkit should be designed and why, I will be happy to have a drink and discuss that matter with you.
Otte: staring into the abyss Posted Jul 28, 2012 7:57 UTC (Sat) by boudewijn (subscriber, #14185) [Link]
I think that you're suffering from something like audience bias. I seem to hear more talk about moving to xfce in these discussions, and that's probably my bias :-)
Otte: staring into the abyss Posted Jul 28, 2012 13:58 UTC (Sat) by augustl (subscriber, #75060) [Link]
Another viable choice is no DE, just a window manager. From what I can tell, this is almost standard in the Arch Linux community.
Otte: staring into the abyss Posted Jul 27, 2012 23:19 UTC (Fri) by drag (subscriber, #31333) [Link]
I hope you take note that the two contenders you pointed out completely eschew the traditional Linux distribution method and are intensely for-profit projects.
(Which I consider a good thing even though Google itself doesn't seem particularly trustworthy. But hey, this is the reason we have open source.)
I still think it is entirely and 100% possible for a true Linux desktop to thrive and attract user and developer attention on a scale that will surpass Apple desktop by a factor of two or ythree. Not that I am picking on Apple or anything.. but just illustrates the size I believe is possible.
Otte: staring into the abyss Posted Jul 29, 2012 20:07 UTC (Sun) by khim (subscriber, #9252) [Link] I hope you take note that the two contenders you pointed out completely eschew the traditional Linux distribution method and are intensely for-profit projects. Well, that's the point. It's obvious that distributions will never agree to provide stable platform for third-party developers and as was discussed many times this is the very first step on the road to the Joe Average. Android and ChromeOS do provide such a platform. Sure, right now it's much poorer then what you have on desktop, but it steadily improves over time - you can not say that about other Linux desktop environments. I still think it is entirely and 100% possible for a true Linux desktop to thrive and attract user and developer attention on a scale that will surpass Apple desktop by a factor of two or ythree. Sadly "true Linux desktop" does not care about being usable platform. It's built around the idea that applications must be adopted/recompiled for the particular flavor of desktop. This is what killed "big Unix" and this what kills Linux desktop today. Sure, it' not 100% apples-to-apples comparison: "big Unix" vendors tried to make sure it'll be hard to port application from one flavor to another while Linux desktop environments favor cooperation. Still ISVs don't want to play these games at all: they want to build an application once and sell it many times - direct opposite of "true Linux desktop". And without ISVs "true Linux desktop" have no chance in hell.
Otte: staring into the abyss Posted Jul 29, 2012 21:20 UTC (Sun) by bronson (subscriber, #4806) [Link]
Not if Linux desktop projects shed users as quickly as they attract them. Sadly, the two big ones enjoy doing this regularly.
Once the DEs realize that most people just want to work without having to relearn everything every five years, then Linux might have a chance on the desktop. (Microsoft is about to learn this the hard way itself...)
Comparing to Apple is exactly right. They slip only two or three invasive changes into each release, and include config items to set things back to the way they were. After five years, most people are happy with the defaults and the config items go away. It's very well done.
I just wish a leading Linux DE would start thinking long term like this. Changing everything all at once is never OK.
Otte: staring into the abyss Posted Aug 3, 2012 6:21 UTC (Fri) by khim (subscriber, #9252) [Link] I just wish a leading Linux DE would start thinking long term like this. Changing everything all at once is never OK. But Apple did that too! But this was done in similar fashion: for many years Macs were sold with two operation systems—MacOS 8 (later MacOS 9) and MacOS X. First MacOS 8 was a default, then MacOS X become a default, but it still included full-blown MacOS Classic emulator and only years later MacOS Classic was finally removed. This is transition from one OS to another, radically different one. GNOME kept a lot of plumbing around yet still broke everything on "one go". Microsoft is about to learn this the hard way itself... I don't think they'll do. It looks like they've lost this capability. The fact that they want to repeat cavalier WP7 switch (similar to GNOME3 switch) with WP8 switch shows that they lost all the people who understood why "redo everything with a flag-day switchover" is a single worst strategic mistake. Sometimes total rewrite produces good end result (the very example Joel uses shows that because it gave us Firefox), but this basically kills the project and starts new one in it's place. Again: look on the MacOS Classic to MacOS X transition if you want to see how to do such switchover right. Heck, look on DOS to Windows and Windows Classic to Windows XP+ transtion (where Windows-on-top-of-DOS was kept around for years while people slowly migrated to new, Windows NT-based architecture). It's really surprising to see disaster Microsoft makes with these transitions today: people who are driving Microsoft today were there back when Microsoft knew how to do that (including Ballmer). Have they forgotten everything? Have they gone mad? What exactly happened?
|
|||||||||||||