Calligra Under The Hood

Just back from Fosdem where I gave a presentation with the above title. It was taped, but I don’t think it’s edited an up yet, but you can get the slides here. I think the talk went reasonably well, but since the organizers had asked for technical content, my content was very technical. Either that, or the timeslot (Sunday 16:00) meant that there were very few people in Jansons and that the people who were there couldn’t come up with any questions. I still do think that the subject was pretty important: how to use the Calligra office engine in your own applications, no matter the form factor your app is intended for.

And development in Calligra is so exciting these days, with 2400 commits since we moved to git. It’s a pity we’re generally speaking too busy to blog or dent. Jan Hambrecht, the Karbon maintainer, has returned and is working on connectors between shapes and clipping. Thorsten Zachmann is properly implementing the long-needed text-on-shape feature. Krita developing is rushing at an enormous pace. Casper Boeman and Sebastian Sauer are rejuvenating the rather unstable, slow and tangled code of the layout engine. Yue Liu is making a go of the development of Flow (formerly known as Kivio). We really should blog more often…

The opening keynote by Eben Moglen was very inspiring, but not everybody could attend since the auditorium was packed while there will still people waiting outside. Eben Moglen called on all of us to make software that promotes freedom on the ground — referring to the current situation in Egypt as well as to the datamining monopoly the United States has achieved in the past decade. Facebook is not harmless, it’s dangerous. Even though they sponsored Fosdem…

It made me think about where Krita fits in there, but the answer is simple: artists who create art to support the struggle for freedom deserve to be able to do that with free tools, tools that cannot be taken away from them. So with my conscience salved, I plunged into the crowded corridors to find the KDE stand where Irina was helping out again this year, despite still being tired from her hospital adventure end of December. Met up with a lot of people, too. I didn’t go to many presentations apart from the keynote and the office track, but it looked like most rooms were packed anway.

The Fosdem organizers really took great care of us main track speakers, with a nice hotel for self and spouse (no free internet at the hotel, though!), taxi between hotel and venue, t-shirt, snack-bag and so on. Wonderful work guys and gals!

All in all, a chaotic, rousing, inspiring and tiring experience. It’s nearly 11:00 now, and I still need to shower. But at least most of my mail backlog is done!

KDE — and proud of it!

KDE — and proud of it!

I think Harri somehow made a mistake in his recent blog on K* == Bad. The Calligra community isn’t moving away from KDE at all. We’re also not pre-empting the KDE move to git — we’re using the excellent KDE infrastructure for hosting git projects. Not only that, but KO GmbH, the company founded by some KOffice community members, actually sponsored the conversion of KOffice from subversion to git. And we’re sponsoring the conversion of the KDE kdelibs and kdebase module as well. And Marijn, the Calligra Tables maintainer, he is also maintaining the KDE kdelibs packages for MeeGo.

Sure, parts of Calligra are used in FreOffice, the only free mobile office suite in existence — but still depends on KDE technologies, because Calligra depends on KDE. And that’s not changed at all from the way it was in the KOffice days.

Calligra uses KDE’s project infrastructure, mailing lists, bugzilla, reviewboard, forums — everything. Only the website is on a separate server (the same as Amarok), just like the KOffice website was, and the release cycle is differnet — as it always was. We’re in the KDE community, many of us are KDE e.V. members…

We’re part of KDE. And proud of it!

Helping KDE move to git

KDE’s migration to git has been a long time in the making, with the first plans being discussed in 2009. One big piece missing has been the actual conversion rules, rules that take into account the history of the KDE software throughout its history. KO GmbH has sponsored Ian Monroe to finish the rules for KOffice and perform the conversion and integration into KDE’s git infrastructure. The final conversion will happen tonight, and tomorrow development will continue using git…

But that’s not all: we’re really happy with Ian’s work, and when KO’s Marijn Kruisselbrink took up maintainership of the KDE mobile profile packages for MeeGo we saw we had an opportunity to take another step: KO GmbH decided to sponsor Ian to convert kdelibs and kdebase as well, and move development of those modules to Git.

So, with a target date of December 20th, Ian is working really hard to make sure that kdelibs and kdebase end up as git projects. Kdelibs will be one repository, minus some kwrite/kate stuff (most likely) and kdebase will become three repos. The first test repo has already landed!

We’re a really small company, so this quite a big deal for us, but we’re very glad we can do this to help KDE move its core platform development to git!


Yesterday I suddenly realized that KDE4 is getting really comfortable to work in, even on my low-resolution (1024×768) laptop. Sure, I have to tweak a bit: all fonts are too big, the Oxygen colours are a bit too colourful, notify sounds need to be disabled, wallpaper changed. But that’s simply the desktop equivalent of moving into a new house. I think I’ve got KDE4 configured just right now.

There are things that make me so totally go wow: the cover flow window
switcher, the breath-taking new login splash, the panel resizer thingy… Krunner is much more useful and much easier to use than the old minicli. There are things that are a bit ho-hum: systemsettings doesn’t do much for me, although it’s comfy enough. I have to run KDE3’s kpowersave to make my laptop sleep when I close the lid. And I don’t like the extra space between lines, especially when
using a monospace font like the old Misc Console. Akregator hangs very quickly.

But worst of all: I have broken Krita. The tools don’t get activated anymore.

But apart from that, I’m quite comfortable now.

Meeting in Nijmegen

Today was my last day at Tryllian, and since my boss had given me the day off I was free to go to Nijmegen, where we had the inaugural meeting of the KDE Research Working Group. As a special guest we had Ralph Fiergolla from the European Commission’s FP7 directorate. Sebastian and Adriaan hosted the meeting in the SQO-OSS project room.

The meeting was quite productive: we decided on an action plan, came up with a couple of likely proposals, had an enlightening presentation by Ralph about the way the EC funding process works, and afterwards a good discussion on how a loosely-knit community of volunteers like KDE could fit into that. The EC is actively looking for project partners outside the group of “usual suspects”, and well, that’s where we want to fit in.

Oh — and Adriaan is dead wrong. There is a place in the Netherlands where you can have truly excellent food for a decent price. In Nijmegen, no less. Indian restaurant Gandhi, Hertogstraat 23. I have never had Indian food  that was that good. We had a couple of dishes I hadn’t seen before, like lamb pasanda (sweet, with almonds). The spices were exactly right. The texture of the eggplant in the baigan bhajee was amazing: soft with a little bite but nothing of the chalk-on-blackboard squeaking that eggplant often has, while not being overcooked. Completely amazing!

Experts wanted

One thing that Ralph Fiergolla made very clear today in Nijmegen is that the European Commission really would like more experts from the free software community to evaluate research proposals. You don’t need a PhD (I haven’t got one either), but you do need a certain amount of technical knowledge, a certain amount of experience so you know what things are likely to work and what things aren’t — but if you’ve been coding for a half a decade or more and have been participating in free software communities for about as long, you might very well be the kind of person the European  Commission needs for its evaluation process.

From my own experience I can say that it’s a fun thing to do, not extremely well remunerated, but you get to meet really interesting people. The evaluation process itself is really thorough and well-designed, and it’s a good thing to have experienced. Sure, there are inconveniences, like having to read about 600 pages of dry prose and doing your real best to really understand the issues and the subject. But what’s a little inconvenience in a righteous cause? And you have to read up on the relevant EU documents, too, which is a bit of work. Perhaps track the rss feed too.

But if you think you have the right stuff in you, why not go to Cordis website, and check it out, and maybe even register as an expert? The KDE community really is full of excellence — let’s make sure that the wider world benefits from that, too.


Last week I was almost completely absent and incommunicado. But it was for a good cause: I was invited to Brussels by the European Commission to evaluate project proposals for the Framework Programme 7, strategic objective 4.2 “Intelligent Content and Semantics”.

I was only one of more than seventy experts evaluating these proposals. The experts were divided into teams of six experts, and we had seven or eight proposals to read, and then to discuss in varying groups of three. This is the building where we were working:

During the evaluation we were held incommunicado: no phones, no network — just about 600 pages to read, evaluate and discuss. This was helped by the fact that the building had been stripped of every outside connection. But taking laptops and mobile phones was forbidden, too. No outside influence, and no taking out of information from proposals.

The procedure was very thorough. Much more thorough than most proposals, actually. I mean — if you’re going to ask for 2 to 4 million euros, the least you can do is run the spell-checker over your document. And probing through the common outer crust of badly edited, sloppy English, most proposals weren’t that good. Some were outright bad, some were decentish, but very few were really outstanding. So it’s not a bad thing that there isn’t money to fund every proposal: we were able to be very selective.

My personal notes on what makes a good proposal:

  • Edit by a native speaker, spell check
  • Give all relevant background information, evaluating experts cannot google.
  • Don’t expound on a grand vision and then forget to actually describe how that vision is going to be implemented
  • Know your area: a rehash of the state-of-the-art anno 1970 is not going to be funded
  • Be really on topic, don’t try to dress mutton as lamb, i.e, don’t recycle a proposal for a different topic with only a little bit of topicality sauce.
  • Don’t plan to build in three years what’s going to be commercially available in six months
  • Try to avoid hiding the inner emptiness of the proposal with buzzwords: it doesn’t work and is a waste of your time and the experts will see through it.

In general, it’s really, really hard to sneak anything past the experts — I was very impressed by my colleague reviewers. So, all in all, a very educating experience, and one that I hope to be in a position to repeat in the future.