SVU's blog

So long

Now that GNOME 3.6 is around the corner, it is time for me to bid adieu. I was committing my code to GNOME CVS/Subversion/git for many years, since 2.5.x series (December 2003) – and I am not planning to do it any more.

For 3.6, the entire keyboard input infrastructure was redesigned. That huge work is done by Rui Matos – he is a real master, no sarcasm attached. Well, in many aspects, new solution has serious regressions. But in many aspects, it is an enourmous leap forward. It depends on what you used to use and what you missed in 3.4 and before. I hope that 3.8 or 3.10 will return the functionality GNOME lost in 3.6, for people who want just working and flexible XKB-based keyboard configuration (like myself). At least with gnome tweaks or smth.

Anyway, it is not my code any more – I am not in charge. https://bugzilla.gnome.org/index.cgi?logout=1

It was a good timing (may be, a bit too late). I was never a GNOME 3 person. My vision never went beyond GNOME 2 ways – I just helped to port GNOME 2 XKB bits to make them work in GNOME 3. I never shared GNOME 3 vision. If my memory serves, I never advocated GNOME 3 decisions publicly. In GNOME maillists my attitude could be really considered as trolling – but in reality it was just an expression of regret and frustration from the person who was not ready for the brave new world. I was trying to stand by the values that were obsolete from the decision makers’ POV. Pathetic, I know:) I still like that heap of mix-and-match components with the title GNOME 2. Well, the interfaces were sometimes good, sometimes bad – but at least they were interfaces. GNOME 2 philosophy gave birth to many standards on freedesktop.org. GNOME 3 would never do that, I’m afraid. I never liked that cathedral of GNOME 3. So my departure was just a matter of time.

But now I am leaving without any bitter feelings, almost. GNOME moves forward – I wish it all kinds of success. It just not my way any more. I am going to use GNOME 3 – not because I like it, just because I am used to it, somehow – and I am lazy to change (and for me it just “sucks less”). Of course, that is unless I’ll find 3.6 totally unusable…

Now, about technical bits.

Libgnomekbd is a dead man walking – the only useful thing is the keyboard drawing widget, that should be incorporated into gnome-control-center in 3.8.

Outside of GNOME, xkeyboard-config and libxklavier remain under my supervision, nothing changes in that area. I am even going to celebrate tonight with new xkeyboard-config release! libxklavier was created to be used by GNOME and other DEs – but not used that way anymore. It is still remains the reference implementation of the xkeyboard-config registry parser, with gobject-introspection – i.e usable in scripting languages (and I know some people do use it with python).

I am going apply for changing my status in Gnome Foundation to Emeritus membership.

Thanks for the good run, GNOME! Most of the time I was proud to be a part of you. GUADEC conferences (Dublin, Birmingham) were quite enjoyable. I learned a lot from your code and your people.

Good luck!

Survey, shmurvey

Here you can see it again. People at Phoronix ranting against GNOME3. People on d-d-l staying kinda cool and professional – and critic-prone. The very same scenario you could see on Slashdot and linux.org.ru and other FOSS-related sites.

Even though GNOME 3 is already 3.2, these battles keep happening. Why? And we seen them back in the days of GNOME 1 –> GNOME 2 transition. Why?

Look at the lessons explained by Guy Kawasaki in his blog. In particular, lesson #2. “Customers cannot tell you what they need”. The same way late Jobs was “the pioneer of the computer as a jail made cool” (c), Gawasaki’s lessons make active marketing look cool.  And yes, “active marketing” is a term with extremely negative ethical value (same as “jail”), as far as I am concerned. What is it about? Inventing new needs for the people. Telling people what they want instead of giving them what they think they want. The results are high profits. Kawasaki says this is the only way to innovate and be commercially successful at that. He is right, most probably. I consider it pushing the innovations down customers throats. That is the thing that made Apple great and rich. That is one of the things that made many people angry with Apple.

It is a common place to accuse GNOME  of following Apple, spiritually, technically, visually. There are funny and serious sides about that. But it seems that lesson #2 was taken seriously by GNOME, when GNOME 3 was conceived. And that is the most pronounced PR issue with GNOME 3 – the way it is perceived by many people.

How many? That is the question of the day. Noone can say. Really. The way the survey is performed by Phoronix is somewhat provocative and questionable – at least it is seriously challenged by the d-d-l folks on those grounds. With all that, the question still does exist. And that question is strategically important.

So, would GNOME Foundation consider performing “proper” survey of what people think about the way GNOME goes? The survey that would be accepted by GNOME developers as useful, reasonable, answering important questions. The survey that (potentially) could help reconsidering the values and goals. Or, alternatively, the survey that would silent (forever!) all people who just want “faster MS-DOS” – by displaying them the real demand in innovative interfaces. “Look guys, there are only 10 of your kind!”

I guess the question is important enough to consider it first priority for the GNOME future, immediate and long term. Would you be able to face the statistical truth?

Enough is enough

I waited several month for Ubuntu 11.04 to stabilize on my good old Power G5. No way in hell. Even X does not start properly. Had to spend several months in MacOS :(( So… considering that I have no interest whatsoever in Unity and other heavily patched gnome components of Ubuntu …  It is Debian now, testing + gnome 3 from experimental. Well, gnome 3 starts in fallback mode only so far – but at least that’s better than absolutely non-working X.

On my laptop, virtualbox for some reason went to disagreement with Ubuntu (inside) as well, so now I have FC15 inside VB, with nearly vanilla gnome3 (working properly with gnome-shell!).

I was some kind of fan of Ubuntu. I am leaving that camp for now. Political reasons, personal reasons, technical reasons… Perhaps, for me 10.10 was the finest hour of Canonical. So long.

libxklavier: good bye, CVS

libxklavier was originally hosted in CVS on sourceforge.net. At some point, it was moved to CVS hosted on freedesktop.org.

CVS is out of fashion for quite a while – but I did not see any reason to switch to git (standard de-facto for freedesktop.org). Now, I made the move. Here it is: git+ssh://git.freedesktop.org/git/libxklavier. All future changes are going to be made in git. I just made last commit to CVS. So long, old friend!

Where is GNOME2? GNOME2 is in EXDE!

I knew it was a matter of time till somebody would decide to continue development of GNOME2, “tried and true”. Here it is, EXDE. There was a number of GNOME forks before – none of them had anything close to some form of real life. I wish this project all the luck – just because too many people find GNOME3 … well … “too different to be acceptable” (putting it mildly).

And, above all, I hope that some good ideas from EXDE (if they are going to happen) could be fed upstream. That would be win-win scenario.

We’ll see…

xkeyboard-config: last developments

I was not blogging here for a while. But the project is still live.

The major event for all of us is going to be GNOME 3 release. There was a lot of changes in the keyboard configuration GUI (now it is search-based, wow!). There was a mockup of what the Regional/Language configuration could look like – and GNOME is very close to it. In order to facilitate that change, some fixes in xk-c would have to be done. Most important, the layout/variant descriptions have to be changed.

In GNOME2, the full variant description was composed of the layout description (usually country name) and the variant description itself. For example, “USA – International”. That schema does not allow creating variants like “Engish (US)”.

Now, the descriptions for variants are full and self-contained. That allows putting there any lines, for example “English (US)”. That is a better approach – but it requires someone to take the latest base.xml.in, walk through it and fix the descriptions. to make them user-friendly If this is done before GNOME3 release – I’ll try to make extra (out of schedule) release of xkeyboard-config. Volunteers, anyone?

PS Last couple of days I got another set of traditional enjoyment, related to Crimean Tartar variant. Here and here. The guy does not get something important…

Keyboard -> Region configuration, GNOME3 style

Bastien did a really great job adopting the kbd configuration panel for the GNOME3 g-c-c style. Thank you, hadess! Even though the decision to drop the model is still questionable for me (discussing in d-d-l), overall it looks more consistent with other g-c-c panels than 2 days ago. Also, a number of bugs related to the keyboard rendering are closed (again, my gratitude to Bastien for filing them).

The future challenge for me is to implement “search” for layouts…

On Ubuntu, Unity and GNOME 3

I think there are some lessons gnome should learn from the fact that Ubuntu effectively quits gnome 3 ride (and bids gnome-shell farewell). It is not only about Canonical. It is about GNOME as well. Sapienti sat.

xmodmap and gnome

GNOME (gnome-settings-daemon) was supporting custom xmodmap files for ages – as a convenient way to tweak the kbd config. I heard several times about people using that feature – even though it was never important, used by minority. Yesterday that feature has gone from g-s-d.

Since I feel that at least rudimentary xmodmap support is necessary, I made libgnomekbd load $HOME/.xmodmap if it exists. Hope it won’t be much trouble for people to change their configs.

Actually, I would be interested to hear here in comments about the ways people use xmodmap with gnome.

The discussion on IRC was quite hot. My apologies for some bad words, lads – I did not control myself well enough. Even though that does not eliminate my points about importance of xmodmap – and in general about our (GNOME) attitude to features used by minorities. Minorities matter. 99% = 99*1%

stracciatella

Do you speak Italian? I do not. Anyway, stracciatella is “vanilla-flavoured ice-cream with chocolate chips”. Ubuntu introduced that idea of nearly-vanilla GNOME quite a while ago (http://www.piware.de/2009/02/the-stracciatella-gnome-session/) but only today I tried a taste of it, with 10.10. Brrrr. Nonsense. Hardly usable (try to finish the session, for example:). A bit too much of “chocolate chips”. If you look at the Canonical discussions (https://wiki.ubuntu.com/DesktopTeam/Specs/Jaunty/StracciatellaSession), there is one essential phrase: “This works on component granularity, not on the level of patches.” Very sensible remark, technically – and honest statement, too. But at the same time, it makes the whole idea more and more useless, as time goes by – because the size of patches applied to standard components grows with every release of Ubuntu. The result is something half-baked that I would not recommend to consume without 0.5 of vodka (I hope the original stracciatella is not that bad).

Whatever tensions are between Ubuntu and GNOME – I do not think it is fair from Canonical to make that kind of bad PR (feels more like FUD) by introducing “NEARLY vanilla” Gnome desktop.

PS I filed a bug in relation to g-s-d, closing it could add a bit more vanilla.