Not logged in
Log in now
Create an account
Subscribe to LWN
Pencil, Pencil, and Pencil
Dividing the Linux desktop
LWN.net Weekly Edition for June 13, 2013
A report from pgCon 2013
Little things that matter in language design
And so on. This is something I have heard previously... Oh, yes. It was when KDE4 was released. It's all just poor excuses, if you ask me.
> I don't see regressions [...]; just stuff that could work better.
Yeah, right. Microsoft said exactly the same about Windows ME and Vista.
> no data on how many are leaving
And, does it matter? Let me ask you something: you really believe that there are more users coming or leaving GNOME?
Why GNOME refugees love Xfce (Register)
Posted Nov 12, 2011 10:44 UTC (Sat) by ovitters (subscriber, #27950)
However, there is a big change and comparing such a big change with the 16th release of the 2.x cycle, and not with 2.0 for me is very logical that it doesn't work out. I see 3.0 as really good and way better quality than I expected.
You're comparing to KDE 4, bringing up ME, and suggesting things. Then stating that data collection doesn't matter. With any big change, there is pain. You have loads of articles written about the effects of change, etc.
We've had pretty much the same feedback during GNOME 2.0.
Without any data, no idea how 3.0 compares to 2.0.
Posted Nov 12, 2011 10:49 UTC (Sat) by dlang (✭ supporter ✭, #313)
unfortunately the press release announcements didn't make this prominent enough (the info was there, but buried). Many of the articles that were then published completely dropped this part of it.
it's not a matter of them "renaming it to a development release"
Posted Nov 12, 2011 13:47 UTC (Sat) by ovitters (subscriber, #27950)
And what I mean with liking GNOME 3.0 is that I consider it really stable and not a development release. But it needs some time before it really shows off the ideas. Meaning with Contacts, Documents, Boxes (3.4), crash app, etc (see http://afaikblog.wordpress.com/2011/11/10/gnome-design-up...). At the moment the extra applications are pretty simplistic. The plan is to launch an beta extensions website on Dec 1st. So finally something that (IMO) blows away the tweaking possibilities of GNOME 2.x. I do think 3.0 and everything that followed is the right timeframe though, it is useful to feedback on 3.0, not 3.0 with hundreds of extensions.
Could've also not released 3.0 and 2.32 (2.32 was only released because 3.0 was not good enough at that time), kept developing without releasing anything for a few years (e.g. not everyone agreed to doing a 2.32; it required time that people rather spent on 3.0). I think "release early, release often" dance is good. That is how you can still change things. If you change fundamentals late, it is just very hard and requires loads of time to adjust.
Also saw some notion that distributions switched too early, but most non-rolling release distributions seem to only provide GNOME 3 as of 3.2. Exception is Fedora, but that distribution is well known to give the latest technology asap. I run Mageia myself, and the latest stable (v1) provides GNOME 2.32 (IIRC); v2 will have 3.4 (+systemd:). Mandriva completely got rid of most of their packages and only support KDE. Opensuse only switched with 3.2. Seems like a normal acceptance cycle.
Regarding forcing, we've ported metacity, gnome-panel at 3.0 and gnome-applets as of 3.2. Various big changes occurred during 3.0 (simplification of System Settings and put the rest in a tweak tool), but I saw that maintainers wanted to do that for years. Only by announcing 3.0 maintainers saw it as a go-ahead (so they first simplified System Settings, which triggered the start of gnome-tweak-tool). But the maintainers could've done the same during any 2.x release. During 2.x it was delayed because someone had to write a tweak tool; in 3.0 it was done and as a result someone wrote the tweak tool.
But oh well, I still use "spatial mode" a bit :P
Posted Nov 12, 2011 20:14 UTC (Sat) by jcm (subscriber, #18262)
GNOME 3 was released far too soon. That would have been ok, had things like the Shell been optional and the existing 2.x real panel and so on stayed around. But instead, everything was thrown away in one go and replaced with an early release. A few years down the line, those of us who felt forced into switching might not be nearly as annoyed once the features catch up with 2.x. As it is, the mindset to me said "you'll take this and like it" (fine with a brand new project, not ok after ten years of using something).
As I said, GNOME 3 finally made me realize that I should not rely on GNOME to remain a consistent UI in the future. It happened to do so in the past, but is willing to throw caution to the wind and force me to change my entire workflow on a whim. This won't do. And I can't use it again :)
Posted Nov 12, 2011 22:04 UTC (Sat) by dlang (✭ supporter ✭, #313)
this doesn't go well with people who thing that a new release should include revolutionary things or it's a 'worthless' release.
it is actually harder to do major changes with evolutionary releases, but the result is usually better.
We see the exact same issue in patches. It's easier to do one 'big bang' patch that changes everything and rips out an old system to put in your new wiz=bang solution.
but if you can go back and create a patch series that is evolutionary instead, it's much easier to get review and feedback.
In some cases you find that the users really didn't want the wiz-bang feature the way you initially envisioned it, but the evolution allows them to benifit from your work even if the end result ends up a bit different.
In other cases the mere act of going back and breaking the change down into logical steps ends up showing the developer places that things can be cleaned up, simplified, generalized, or otherwise improved that were not obvious in the 'big bang' change.
Posted Nov 12, 2011 23:18 UTC (Sat) by jcm (subscriber, #18262)
Posted Nov 15, 2011 2:39 UTC (Tue) by nevets (subscriber, #11875)
Look at how Linus changed Linux 2.6 into Linux 3.0. It was no different than any other release. There was absolutely *no* wiz bang features to merit a 3.0 (except that ftrace function tracing was redesigned and now works with modules :-).
I've been using Linux since 1996, and played around with fvwm at first, and I forgot when I started using gnome. I liked it at the time. IIRC, when gnome 2 came out, it was radically different too, and lots of things broke. I may have switched away from gnome 2 at that time, but things were not as developed back then so radically different wasn't as different as things are today.
When gnome 2 settled down (and brought back a lot of features that it removed), I started using it again and its taken me 10 years to perfect a workflow. I've tried many, and what I ended up with was something that works great for me. Actually, I really only use gnome panel, as I've dumped metacity the first day it came out. I'm a sawfish lover, and when that is hard to install, I've actually liked xfwm4 (I've started using that in the last year).
Last week I did an update to my main box (debian/testing) and it blew away gnome2 and installed gnome3. The gnome-panel has none of the functionality I've come to depend on. My 10 years of perfecting a workflow just went out the window. I bitched like hell, and the only thing that I hear from the damn gnome3 lovers is "oh, its better if you do it this way". BULLSHIT! I've spent 10 years perfecting something to get my work done and I'm not about to change everything just because people like eye candy.
gnome is just a platform to get work done. If I notice it, then it's broken. The same goes with kernels and computers in general. You should not be focusing on the platform that you are working on, you should be focusing on your work. The platform is the tool for your work, not the work itself.
This is the heart of the problem with gnome3 developers. They are focused so much on gnome being the end product, and force users to do it their way.
The reason I started using Linux in the first place is because it let me control the computer, unlike Windows and Apple which make the computer control you. I have a strong feeling that gnome is trying to control me and trying hard to keep me from controlling gnome. This is why I'm so pissed off, and yes, I have started setting up my desktop with Xfce.
Posted Nov 15, 2011 8:06 UTC (Tue) by deepfire (subscriber, #26138)
Time is spent in this process -- months, years.
Gnome 3 wasted all these efforts, with the promise of something better.
Something was indeed better, but still, the amount lost was far too much to be acceptable.
So, in the end, the most damage inflicted was to the people who invested heavily in workflow adaptation/customisations.
Copyright © 2013, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds