Keyboard indicator in Ubuntu 10.10: Disclaimer
August 20, 2010 3:07 pm g-s-d, gnome, libgnomekbd, ubuntuTWIMC. The keyboard indicator to be shipped with Ubuntu 10.10 has little to do with the standard gnome keyboard layout indicator. Canonical moved the thing to libappindicator library which was not officially adopted by GNOME.
So, there are some important consequences:
- If you are end-user running ubuntu 10.10 and you have complains about gnome keyboard indicator, you’re not really welcome to bugzilla.gnome.org. First, you should file your problem at Launchpad. If the Ubuntu team finds the problem down the stack, they sure will escalate it.
- A lot of Ubuntu users (including myself) were happy to utilize non-documented but working support for flags in GNOME. In Ubuntu’s implementation that feature is not supported, AFAIK. You’re welcome to complain in Launchpad.
- Single-click layout switching does not work any more (I was told). You’re welcome to complain in Launchpad
As a GNOME developer, I am very concerned about the fact that Ubuntu is starting to create its own version of GNOME. All distrovendors have patches, but Canonical seems to be gone a bit too far, on my taste.
As an end-user, I am quite irritated. I am thinking about switching back to Debian or Fedora (considering the fact that Ubuntu’s support for PowerPC is not great anyway).
I would appreciate if someone provides explanation on how to get original gnome keyboard indicator in Ubuntu 10.10. I did not install that version yet – but people are asking.
August 20th, 2010 at 3:36 pm
[…] This post was mentioned on Twitter by Zuissi, Larry King. Larry King said: SVU's blog » Blog Archive » Keyboard indicator in Ubuntu 10.10 … http://bit.ly/aSBl4N #Ubuntu […]
August 20th, 2010 at 3:38 pm
[…] Sergey Udaltsov […]
August 20th, 2010 at 4:01 pm
Sergey, for exactly the same reasons i already moved to debian 3-4 month ago.
Regarding the vanilla kbd indicator i suggest you to create your own ppa with a packages you need.
August 20th, 2010 at 4:34 pm
The question that GNOME as a project has to answer is how far can a vendor differentiate and still call it GNOME?
Meego and Litl are creating vastly differentiated interfaces built on top of a lot of the same framework that GNOME’s interface is using..but they don’t make any claim that its a “GNOME” desktop. So people are fine with the differentiated work they are doing.
The question is how far can a vendor go to extend/experiment/change what GNOME is and still be able to call it GNOME? Until GNOME as a project can provide guidance on that question vendors can pretty much call whatever they ship GNOME and all it will do is cause frustration like what you are expressing and brand dilution for GNOME as a project.
-jef
August 20th, 2010 at 4:39 pm
isn’t this like ximian’s desktop?
August 20th, 2010 at 5:16 pm
All distrovendors have patches, but Canonical’s aren’t accepted.
August 20th, 2010 at 5:17 pm
Given upstream Gnome hasn’t accepted libindicator (yet), how do you think this should be handled instead?
You can take a look at the downstream changes over here:
https://code.edge.launchpad.net/~ubuntu-branches/ubuntu/maverick/gnome-settings-daemon/maverick
The patch that adds your stuff, I believe, is ./debian/patches/06_use_application_indicator.patch
Why the way, welcome to the indicator applet! Your software is consistent with everything around it, now, and it looks pretty nice:
http://lucidfox.org/posts/view/607
August 20th, 2010 at 5:20 pm
[…] more here: SVU's blog » Blog Archive » Keyboard indicator in Ubuntu 10.10 … Рубрика: Разные рубрики | Метки: a-bit-too, bit-too, create-its, […]
August 20th, 2010 at 6:07 pm
For similar reason switched to Fedora. They are going too far. It’s disrespectful to upstream.
August 20th, 2010 at 7:41 pm
For a few versions now Ubuntu has had the package “gnome-stracciatella-session” which installs a session selectable from GDM which is vanilla GNOME, without Canonical’s changes. I don’t remember what it *currently* does, but I know one of the changes is that it uses notification-daemon instead of notify-osd. I *think* it also removes the indicators and MeMenu stuff.
August 20th, 2010 at 9:07 pm
While using “Stracciatella” might be an easy way of using “mostly vanilla (Gnome), but with some brown chocolate chips”, the problem still remains. No matter what Jono Bacon tries to tell you, Ubuntu is forking Gnome.
I don’t think the problem is just that Canonical is trying to differentiate Ubuntu at the wrong level*, there’s something wrong with the way upstream Gnome is managed, too, if they really feel the need to fork it. As a Friend of Gnome, I hope that Canonical/Ubuntu can become a productive member of the Gnome community, rather than the producer of a somewhat unfriendly fork it currently, and increasingly, seems to be.
(If there was another Debian-based distro which offered Gnome as a first class DE and had releases synced with Gnome, I would most probably switch as well.)
* I’m not sure about the numbers, but I think Ubuntu doesn’t really need to differentiate from other distros anymore, it has a user base which is probably bigger than the rest of them put together
August 20th, 2010 at 11:01 pm
For what it is worth, I have a set of guidelines at
http://fedoraproject.org/wiki/Staying_close_to_upstream_projects
It is a general policy within Fedora that we must not deviate from upstream as much as possible.
August 20th, 2010 at 11:33 pm
Oh, the horror! Somebody is doing something with my code!!!1!
=))
August 21st, 2010 at 12:22 am
+1 with Mircea.
same people who would complain later “hey they don’t contribute back to gnome”.
August 21st, 2010 at 12:36 am
free software.
everybody can get it, modify it and distribuite it.
i can’t see the problem.
August 21st, 2010 at 4:13 am
Perhaps I’m reading this wrong, but are you bothered that somebody has made their own changes to GPL’d code? Isn’t that partly why open source exists in the first place?
August 21st, 2010 at 7:47 am
Is Fedora really going to be that much better for PPC? Fedora has been consistent over the years about saying PPC support is second tier – they are not going to invest resources into maintaining it and it is entirely a volunteer effort. If you’re PPC and are going to switch I think Debian (or perhaps Gentoo) would be a better bet.
August 21st, 2010 at 9:04 am
I think that Ubuntu have no easy way to push some changes out of the sudden. I’m not an Ubuntu fan of this policy but as far as I know: Beagle, XGL and Compiz when they were appear suddenly, they made eventually in upstream (at least some projects recreated them like Tracker or AIGLX, and others become the default WM for composite desktop as Compiz). In a similar fashion many things if they were just published upstream, will make the those things to happen to a slower rate. Also forking, if you can afford it, is in the F/OSS spirit, it’s eventually about the freedom to do whatever you want to do with this code (the main advantage of being open). Ubuntu did it’s Upstart, and on a similar fashion, SystemD shows that not everything could be done “upstream”. I think that is at large an issue of: we want to fork to test our this new concept and if it will eventually get better than the upstream, the upstream will take it as the default, or remake a similar implementation.
August 21st, 2010 at 9:45 am
I’m personally more concerned with regular distro patches. Looking at Fedora’s GNOME patches makes their upstream-upstream-upstream motto seem like greenwashing.
August 21st, 2010 at 11:06 am
That’s the reason I switched to Fedora starting from FC10. Canonical is the only company not working with upstream at all. Like somebody said, I’d rather they call their desktop an other name.
I was really happy with the success of Ubuntu, but now I’m not at all anymore. It’s not jalousy : as a developer, I really can not stand their position.
As I was using Ubuntu, I noticed a lot of updates with many patches, with only a few upstreamed later. I can not understand this as a developer : this is not good as it induces more work to maintain the patches over the versions, and because it’s not whealthy for Gnome (the project) at all : it does not improve. Plus all distributions/company must do the same fixes.
Of course, it’s harder to push changes upstream, because the quality is not there sometimes, simply because of time issues and workarounds instead of real long-term fixes/changes. But still, they could try !
Their Gnome upstream contributions are so small (latest I remember is a port of Gdm to upower, ~10 lines of code), they did not even bothered justifying their proposal of appindicator to Gnome, they only asked once, did not answer on the most important questions, and they feel good about it.
No discussion, nothing, the exact opposite of what Gnome/Kde/Xorg people have done : standardizing on a set of features/API/specs through freedesktop. They only tried a bit with libappindicator on xdg but they did not fix the main design issues some developers pointed at.
And like Jono and some Ubunteros saying “the patches are avail. in downstream” : do you really think developers will peek at all the git/bazaar/svn repo of the world ? Really …
That’s the job of a big player, and I simply do not understand why everybody manage to do it but not Canonical.
I hope the current “noise” being propagated by users and developers will make things change for the best.
August 21st, 2010 at 11:38 am
Well, contrary to what people think, Ubuntu is not having more marketshare than the others united.
According to http://blogs.magnatune.com/buckman/2010/03/magnatune-sends-check-to-gnome-foundation-thanks-to-rhythmbox.html , I found that Ubuntu account only for 32% of the gnome desktop who shopped on magnatune, using rhythmbox.
( 7159$ in 39 months, ie 183.5 per month for RB, and for Canonical modified RB , 1017$ on 17 months, so 59.8$ par month, so 1 out of 3 sell come from Ubuntu, in the best case ).
Which basically mean that 2/3 of the Magnatune clients on rhythmbox are using another distribution, or are recompiling rb to remove the Canonical patch. Since RB is the default player for almost all gnome distributions, and since there is other DE than GNOME, and some kde centric distribution, I think the claim that Ubuntu has more marketshare than all united to be false.
On the other hand, I may be completly wrong in my stats, so if someone actually find were I am wrong, I would be happy to hear it.
August 21st, 2010 at 1:41 pm
@Michael : the methodology/sampling you base your stats on is flawed, because we cannot know if the Ubuntu demographics are the same as the Fedora user base demographics. I do not think that people who chose Fedora have similar tastes and ideals than the people who choose Ubuntu. This will influence buying decisions, among other things. FWIW, maybe all your stats tell is that Ubuntu users are less likely to purchase online music.
Also, being the most popular distro around means that your sample size is many times bigger and more diverse. You may have developers, gamers, grandmas, joe plumbers, etc.
You cannot draw any conclusions from the music purchase stats until you have similar samples with other variables factored out… nevermind, even if you did that, it’s still not a valid measurement. It isn’t a census, it’s a purchasing decision measurement.
That’s why Canonical is now implementing a user census phone-home thingy for Ubuntu OEM installations.
Last time I heard, with the fuzzy stats we had, we had over 12 million Ubuntu users.
August 21st, 2010 at 5:25 pm
To those who say they don’t see what the problem is, I quote the Jargon File:
(Free Software allows forking, which is a good thing, because that way a project can be forked if its original maintainers become a problem. With Gnome, that isn’t the case.)
August 22nd, 2010 at 5:02 am
@Jeff
Explain to me how Canonical gets that 12 million userbase estimate. I dare you. Explain it in detail. You can’t sensibly be critical of Micheal’s methodology and then hold up the “fuzzy” estimate of 12 million without telling us how that number is collected. Not only is it not fair, but its offensive.
If you are going to be critical of Micheal’s estimate then you _must_ present the methodology for the 12 million estimate.
The reality is Canonical holds its counting methodology for Ubuntu as a trade-secret and that is just not appropriate for a _community_ project which Ubuntu is supposedly is. It’s also very interesting to note that Canonical has yet to provide an estimate for its Landscape customerbase..a number they don’t have to estimate and know with very exacting specificity. I’ll leave it to you to decide why the executives of a for-profit company would choose to release one number and not the other.
So please.. either explain in detail how the 12 million is created beyond a hand-wavy “boil the numbers down” comment from an Canonical exec… or stop repeating the number as it derived from a valid estimation methodology.
-jef
August 22nd, 2010 at 7:57 am
If RedHat did this it would be development, Canonical do it and it’s forking. They can’t win can they?
All the this anti Canonical/Ubuntu stuff is sour grapes.
August 23rd, 2010 at 4:19 am
I’m not saying the 12 million estimate is right either (thus the use of the word “fuzzy”). Just saying that you can’t draw any conclusions from online music store purchases, two completely different realms.
August 23rd, 2010 at 4:20 am
Furthermore, I am not an ubuntu fanboy or related to Canonical in any way, and I did not intend my comments to be offensive. Just pointing out methodological flaws as requested.
August 23rd, 2010 at 5:34 pm
Jeff,
Every estimate is “fuzzy.” It’s very simple, refrain from quoting the Canonical estimate until there is a publicly communicated methodology for how its derived. Including it as a parting shot without being critical implicitly legitimises it.
-jef
August 23rd, 2010 at 6:04 pm
If you like vanilla GNOME, you probably want to switch to Fedora or something like that. Ubuntu wants to do a lot of things and have a pretty clear and stated direction, one that isn’t accepted by upstream GNOME, and so needs to deviate to evolve. Fedora on the other hand wants to stay as close to GNOME as possible. Both are perfectly reasonable goals, and will suit different people. Especially as long as the underlying framework all stays the same and the foundation is common, I see nothing wrong with either. Though I would like it if GNOME at least was open to some of the patches proposed…
August 26th, 2010 at 12:50 am
Stoffe:
Other than libappindicator and the patches that require its usage… what proposed patches that have been proposed have not been taken up by GNOME?
-jef
August 26th, 2010 at 11:43 am
To all to have doubts about how big Ubuntu’s userbase is. Well, in our country, Philippines, a 3rd world contry, in most cases no one knows Fedora, Debian, Gentoo etc… If you would go to a computer shop and ask if they have computers pre-installed with Linux they automatically assume it’s Ubuntu. For that reason alone, even though that doesn’t mean those who know Ubuntu would use it as their OS, gives us a clear indication that Ubuntu will always have bigger userbase compared to other distros. Why? Because it’s so popular and so EASY to use! Does it ring a bell?
September 22nd, 2010 at 9:29 pm
Ubuntu easy to use? Are you kidding?
Maybe they are easy to use because they could not fix the layout switching in their 8.04 LTS release fo the whole term of support (3 years)?
Or because they did not include a package needed to connect to Internet via VPN (which is needed by 80% users in Russia and CIS) on their CD for more than 3 years (!) rejecting it for 7 consequtive releases dispite it could cost them nothing, all needed packages already were in the repos? So people had to install Windows just to connect Internet and download the needed packages.
This is the related bugreport: https://bugs.launchpad.net/ubuntu/+source/ubuntu-meta/+bug/107738
123 people interested, rejected for 7 releases as a ‘fringe case’.
The same bug in SUSE:
https://bugzilla.novell.com/show_bug.cgi?id=533580
One user interested, fixed just in one month, though this is noit really needed in OpenSUSE because Yast already provides better functionality.