The browser myth

11. October 2010

(In reply to Henri’s post)

Most people will probably answer that the web browser is the application they use most often. That might be true, but is it the application they spent their time with when they are working? I certainly hope not!

So, let’s dive a bit more into the topic and think about what people are doing during their work hours:

  • Acquiring information: web browser, PDF, Image viewer
  • Communication: Mail client, Phone, IM (no, people are not using GMail for their work mail, believe me, they use Outlook or Lotus Notes usually)
  • Creating content/working: Office applications, specialised software (CAD, Software development, publishing, etc.)

I haven’t made any statistics of course but I feel this boils down to about 20% acquiring information, 30% communication and 50% creating content or at least this is probably what your employer wants you to do. Basically that would mean you only spent maximum of 20% of your time with web browsing while of course the web browser might be open in the background all the time.

Windows doesn’t have a 90% market share because it integrates the web but because it has an ecosystem that provides all this other stuff people need to do.

That doesn’t mean we shouldn’t be working on better integration on web applications or the web in general but we must know that our target audience is most likely not the professional desktop. There are reasons why many people have a MacBook but work with Windows in their company. There are reasons why people use their iPhone for web stuff only but still have a Windows PC at home. It is because besides all the cloud discussion taking place nearly nobody is creating content with web applications that isn’t made for the purpose of web publication.

If you have statistics, prove me wrong, please! Maybe the company I am working at is totally unrepresentative…

Update: Many peopple comment that they use GMail and Google Office Apps for their work. I doubt that this is general true. Most people reading my blog (or more likely Planet GNOME) are probably geeks or working in a very computer oriented environment.

However, I didn’t look at what apps I am using most but what the people I work with in an engeering department use most. They usually have about 5-20 windows open at a time where the browser is one of them while the work in Microsoft Word or Excel most of the time.

Maybe there are also big differences between Europe and other parts of the world because people and companys here are caring about privacy and data security a lot and therefore don’t move their important data and mail into the cloud and therefore aren’t using Google Apps.

I also really wonder if you all use the web browser for programming. I have never seen someone using a browser based Eclipse, Emacs or Vim even on very geeky conferences. And if programming is your job I certainly hope this is where you spent most of your time.

If you wonder, there already is a discussion on this bug report.

Today, I got the long awaited approval mail from the board for the hackfest (well, for the travel sponsership for people attending). This means the hackfest will take place as planned from

December 2nd to December 5th 2010

This also means this is your last change to think about attending! All the people listed in the wiki should have received additional information about the next steps. If not or if you have any questions, please contact me!

Andreas Proschofsky made an interview with Jono Bacon that is really worth reading.

Some comments on this from my side:

Actually that’s all upstream work, it’s just that Ayatana is the upstream

Well, it is not. If you customize an upstream project (GNOME) you cannot call your result upstream anymore, it is downstream. And Jono doesn’t mention that Canonical requires copyright-assignment for all of the parts of Ayatana.

Update: My original wording was: “Technically upstream means not done by a distribution (downstream)” which seems to have led to some confusion so I try to clarify that.

Historically a lot of the work has been done in upstream GNOME but that’s changing

Prove? I don’t think it’s changing even if Canonical probably wants to change that.

So sure it’s a delicate situation but it’s the same with Red Hat building GNOME Shell which is a completely different user experience to GNOME.

As I think that Jono knows better I wonder why he is oversimplifying things here. While GNOME Shell is pushed by Red Hat everybody was invited to contribute and there is also a community of people not working for Red Hat that is heavily contributing to the Shell. GNOME Shell is not a Red Hat project.

Part of it is the fact that the design team that we’ve got working on Ubuntu has a different set of ideas. So besides the mobile space – where we are building a new User Experience with Unity – the focus for Ubuntu is GNOME with these additions.

I think everybody in GNOME would love to work together with the Canonical design team (and luckily it is happening sometimes). If they would present different ideas how they see the future of GNOME (Shell) there would be really nice discussion IMHO.


I really don’t mean to say that Canonical is evil because that would be unfair. They support GNOME in various ways with sponsership, patches, bug-reports and ideas.

As they are a company aiming on making money I can kind of understand their strategy on trying to be better, more innovative and different from other Linux distributions. This is their legal goal with Project Ayatana and it is one of the reasons they do not care too much about pushing things upstream.

But I think they make a mistake if they think that they can replace thousands of volunteers working on upstream projects with developers they have to pay in-house. In short term this may lead to more innovation but in the long term they will suffer in quality and quantity.¬† You might remember the story with Novell’s main menu which was also innovative but didn’t ever end up being widely used.

Anyway, it is far more important to look into the future than into the past. This will also mean that people should push upstream work and GNOME Shell on the Ubuntu Developer Summit and I am sure there are many people there who care a lot about GNOME and Ubuntu.

Save the icons

7. August 2010

Stop Vincent from removing these important icons from gnome-desktop module:

Be honest, can a desktop live without these?

So, finally the Python plugin originally developed by Ishan Chattopadhyaya in GSoc 2009 has made it’s way into the master branch after some heavy modifications.?? As usual the auto-completion and calltips are fully asynchronous and won’t be in your way while typing. You need to have the rope libraries installed for it to work (and that’s not checked for now…) and they are used as backend. As I am not a Python programmer I cannot really say if the support is complete but it shows a reasonable amount of information. As python is not a strong-types language this are actually a bit tricky and all the dirty work is done inside rope.

Enough words, some screenshots:

Autocompletion for python objects


Python calltips


Patches welcome…

All Python programmers are encouraged to test this, give feedback (Bugzilla) and ideally write patches to improve this.

(cross-posted on Anjuta News Blog)

GUADEC follow-up

2. August 2010


As others have mentioned, GUADEC was great. Thanks a lot to the organizers. Especially it rather felt like being part of a big family instead of just being a guest there. And I think that was the first GUADEC where people started dancing. So congratulation to Stephane, Lenka and Patricia for winning the dancing awards!

It is also always great that people not yet involved with GNOME come to GUADEC because they want to become a part of the project.


With the discussion about delaying GNOME 3 and after watching the gnome-shell talks I wondered a bit if we go into the right direction. Actually, I think we do though nobody can say today if all this will be successful. My point is that we basically copied the desktop idea from others in the past years with little changes and little innovation. Now we really try to innovate which brings the risk of failure but also the big chance to really increase our market share. Remember that we had this 10×10 goals?

So, as conclusion, we might fail but at least we tried.

Development Documentation and Tools Hackfest

The wiki page has been updated to finally contain a date, which is 2nd to 5th of December 2010. This might still change if it causes problems for too many people, so please contact me if you cannot make it on that weekend. Also we still need sponsorship and I didn’t get an answer from the board, yet. Stay tuned!


While still trying to release a new development version of anjuta I mostly worked on getting the python support from GSoC merged. There is still a lot to polish but the basic stuff is now set up in the python-support branch and will hopefully hit master soon. This will eventually lead to auto-indentation, auto-completion and calltips for Python.

Long story short

On LinuxTag I was talking to Vincent and Stormy about improving our developer experience and Stormy was proposing to do some kind of hackfest. At that time I really wasn’t sure how that should be done and who to attend.

Anyway, after talking to many of the documentation guys at GUADEC we kind of came up with a plan and a basic set of people to attend.

The Goal

In short: Make GNOME an awesome and state-of-the-art development platform. That probably means to redefine lots of things in the developer documentation and tools to attract new developers and make existing more happy.

In the end, that should mean more and better GNOME applications of course.

When, Where, What, Who

The only really fixed thing we have is a locations that is

Openismus Office



Thanks a lot to Openismus for sponsering the venue!

Currently we are thinking of scheduling the hackfest at the beginning of December 2010, detailed dates are to be announced when we know who is attending to avoid conflicts.

I already listed some people interested in attending on the wiki page but we really want more people to attend.

Why you should attend

We are looking for the following kind of people

  • (Development) documentation people
  • Developers of the development tools
  • Developers interested in making things easier
  • People who tried to get started in development and ran into all these difficulties we want to solve
  • Especially people doing Python (Julie: hint, hint!) and Vala development (or some other language) because we really want to make sure to have those as first-class citizens.
  • Designers to make things looking great
  • Marketing people because a lot of what we do will be a great marketing thing

I also know that quite a number of people at the major distributions (Red Hat, Novell, Canonical) are working on fixing that problem and I would love to have them around. We are also still looking for more sponsers to be sure that everybody is able to attend.

Stay tuned, don’t hesitate to ask and if you are thinking you could be the right person to attend, well, you are!

If you look to the left you will see that after nearly four years on Planet GNOME I now have a hackergotchi (thanks Kat) and I am joining the famous hat fraction.


This is even more important because now you can finally find the person that broke half of your development environment at GUADEC because, yeah,
I'm attending GUADEC
and I am

Sponsored by GNOME Foundation

This is also a good place to remind you of the world-famous FreeFA cup where you can save the honour of your country. In the last years this has been lots of fun so be sure that you aren’t missing it.

Five Questions

1)Who are you and what do you do?

Sometimes, I really would want to know that ;) On other days, I am a 25 year old student/trainee who currently builds trains for work and hacks on the GNOME Development Tools in his spare time.

2) How did you get into GNOME?

I wrote a patch

3) Why are you coming to GUADEC?

It always much better to meet all the people you only now from blogs, IRC and mail in person. Most people in the GNOME community are actually pretty awesome and fun to hang out with. And don’t forget about the beer event of course…

4) In 1 sentence, describe what your most favorite recent GNOME project has been. (Doesn’t have to be yours!)

GNOME Shell (yeah, I know some people said this already), because it is the first time since about Windows 95 where someone really tried to reinvent the desktop user experience.

5) Will this be your first time visiting the Netherlands?

No, I have been to Amsterdam and I sailed on IJsselmeer twice. Nice country, a bit flat though.


As Vincent already mentioned we had a nice talk at LinuxTag and there is also a (german) radio interview with me about GNOME3.

Was fun to talk to some people, especially to one of the KDevelop developers who faces about the same problems as we do: There are 10 people working full time on QtCreater while Anjuta and KDevelop probably have 5 spare time contributers each :| Awesome though, that we are still competitive :)

(cross-posted from the Anjuta News Blog)

Today Anjuta 2.31.3 was released today and it brings some nice features and bug fixes (see NEWS)

What’s new?

  • Massimo Cora’ put a lot of energy into the speed of the symbol database population and brough it down to about 15 seconds for 1000 files.
  • Abderrahim Kitouni finished up his GSoc ’08 work and the language support for vala got merged into trunk. This means we now have autocompletion and calltips for Vala.
  • Naba Kumar ported the class-inheritance plugin from GnomeCanvas to FooCanvas, made it much faster and integrated it back into master

What’s happening behind the scenes?

  • S√©bastien Granjoux worked hard on the new project-manager branch but unfortunately it won’t be finished in this cycle. But once it is finished it should bring a much better project management to anjuta.
  • Naba Kumar is working on improving the database queries for the symbol-db which should in the end result in faster and better autocompletion and in general better code.
  • James Liggett started working on a completely new and innovative interface to the git version control system. It will integrate version control much better into the workflow.
  • Abderrahim Kitouni also worked on a plugin loader for python and javascript plugins. In the end we should hopefully support even more plugin languages with gobject-introspection. He blogged about it.
  • In general we follow the development of libpeas with much interest. This might replace our own plugin infrastructure in the future.
  • The glade plugin was cleaned up. It doesn’t provide any new features (actually, some were even removed) but it should be more stable now. This work is targeted to allow to drag signal handlers from glade right into the code. Currently this mostly meant cleaning glade internals but there is still hope finish it for 3.0.

How you can help!

  • Test the latest release and file bugs! Though declared as “unstable” it should be equally stable as 2.30.x.
  • Help us fixing bugs! There is lots of stuff to do including a new icon (we would like to keep the horse which was a unicorn before the horn was ripped off, otherwise you are pretty free) and a new splashscreen.
  • Help to improve language support! We have language support plugins for C/C++, JavaScript and Vala now but most of the Anjuta developers are C coders. We need people to test this and if possible also fix it. Vala support is written in Vala itself which should make it easier for you.
  • Documentation, especially tutorials are another point were you can help us and any potential GNOME contributor in getting started with Anjuta.