ervin's piece of web

This time it’s really over… I’m writing this from home. I’m a bit sad I’ll miss you all! It was really great. My trip home was uneventful which is good news! My luggage has not been lost… pfew!

I noticed that the Kalyxo interview is now online on the dot.

Just for fun I have used the qwertz mapping during aKademy… and now I need to re-learn the azerty keymap it seems. It’s my typos day! =)

I’d like to thank the whole aKademy organization team! You’ve made a really great job for all of us. I’m sure it’s really appreciated.

Ok, my “reportage” from aKademy is now over. My blogging rate will surely drop from now.

Ok, we already feel it’s the end of aKademy here… less and less people hacking in the computer rooms. It’s a bit sad, but it has to end anyway.

I unfortunately won’t have time to hack today since I have to return the laptop I’m currently using. Moreover, I need to go back to the youth hostel this I won’t be in the same room tonight. I’ll find some more time for hacking when I’ll get home.

On another note, Peter Rockai and myself have been interviewed this morning mainly to discuss our work on the Kalyxo project. It should be published on the dot later today. It’s a bit strange for me to be interviewed… it’s the first time I have to do this.

We’re near the end of the aKademy… Tomorrow will be the last day. But today was quite interesting.

I woke up early because I had to be present for a talk… The talk I gave with Peter Rockai about the Kalyxo project. For my first talk in english it was not that bad. I was understandable. Maybe our talk was a bit too technical for a user conference… but since the project is still quite small we don’t have final products to showcase yet. Anyway, some people expressed interest after we gave our talk, it’s a nice feeling.

Then I attended some other talks. The first one about Kommander was interesting, it really seems to be a nice glue technology and I’ll look more closely at it for prototyping tasks. Then, Jonathan Riddell’s talk about Umbrello was interesting too… maybe a bit too UML centric, but it’s nice to see someone working on this. From what I’ve seen it made huge progress.

I’m working with David Faure on the trash:/ ioslave. It’s very interesting and I try to learn some of the KIO black magic he knows. This is an interesting piece of code to work on. Moreover David is really friendly and patient… I appreciate to work with him a lot.

Argh! I still have to sort out some of the design thoughts I have for the new devices:/ ioslave… I need to find some time to have this on paper and discuss it with the relevant persons.

Today, I managed to arrive at 11:00 to the aKademy… I slept enough for this night at least. I’d like to do the same this night, but… since I have a talk at 10:30, so it’s not advised to arrive at 11:00

So the working day was short… I discussed with David Faure a lot in fact. Mostly of hacking kioslaves for the purpose of the devices:/ ioslave. At one point we became slightly off-topic since we share some cultural background.

I started to help David on the future trash:/ ioslave (a least I try). It’s a good thing to address this, because the current trashcan implementation is not that fortunate. I’m pretty confident that this ioslave will improve things a lot for the user.

This blog entry is quite short today… But there’s not much more to say this time.

Today I was not able to find time to hack on anything… Too many Bird of Feather sessions I had to attend.

Started with the KDE 3.4 vs 4.0 meeting… It was quite interesting in my opinion to see the developers not speaking only about coding but also about processes management. Some people I know should really see this because they still believe that free software hackers are basically people spending there coding like monkeys (should I had they are not part of my friends?). Nothing is official yet, so I won’t tell what the result is (or if there’s a result)… but anyway I’m pretty excited to see a new release cycle starting, and lot of great stuffs will happen.

After lunch, I attended the NX BoF… The topic was really interesting. I really appreciated Aaron Seigo plans for this technology, and Matthias Ettrich point of view about it. They are really clever… and I have a strange shy feeling in their presence. I know it’s stupid, and they don’t act as if I had to be inferior or something like this. Strange…

Finally, the KDE Debian Integration BoF started. I was co-chairman with Peter Rockai. I have to admit that I’m a bit disappointed. We surely could do better and we maybe managed to be boring actually. But, we had some interesting input from some of the attendees… so it’s not that bad.

Because of yesterday coding marathon I only woke up at 11:00 today… So it’s now official, I went to aKademy very interested by the usability talks and discussions and managed to miss everything except Aaron’s talk. It’s really a pity. Anyway I hope that those discussions will lead us to a well organized cooperation with real usability expert. We need to find our own process to handle usability aspects in open source software.

Apart from this major failure, I attended the KDE Quality Team BoF. I’m not really a PR skilled person, but the discussions were interesting. An idea was even given by mornfall to allow user to propose “What’s This” texts when they are missing. He started to implement it this evening, and it’s shaping really well.

On the hacking front, I finished the first stage of my work on the devices:/ ioslave. You can now give more convenient name to your devices. But the whole code is a bit old and difficult to extend, so we’ll discuss the new architecture until the end of the week. The new implementation will be done after aKademy.

Now, I’m going to find something else to hack, like helping mornfall on his current work, code for Kivio or maybe help David Faure a bit on the upcoming trash:/ ioslave.

So, yesterday was the first day for the coding marathon. I have not many things to add… I was so tired when I finished that I didn’t want to blog at all.

One particular event, except from the coding, was the PGP Key Signing party… really a game for geeks. You make a little dance around a room, showing you ID card and your key fingerprint. Of course most of the people laugh at your photo ID.

Of course, because of yesterday, I managed to get up too late to attend the first talk of the day about usability. But I was here to listen to Aaron Seigo’s talk… very interesting!

Today, some coding again. I’m working on the devices:/ kioslave. At this rate I hope to implement the planned features before the end of this week.

Another “got up early” day… even harder than the previous one since I slept less! But again a lot of very good talks! It begun with three talks about different pros and cons : MAS, GStreamer and NMM.

Firstly, MAS is interesting but is really specialized to audio… It’s fun to see it in action controlling audio playback on another computer. It seems well adapted to the LTSP project for example.

Then we had a talk about GStreamer with Scott Wheeler (without the technical problems of the first day… which was quite fortunate for him) and Christian Schiller. What is really neat about this project is that it’s able to run on embedded systems. More over they try to promote free multimedia formats which is always a good thing. The only concern I see is about binary compatibility… it is widely used in Gnome, and if it’s adopted for KDE4 they’ll surely have trouble to cope up with the release cycles of both projects. In particular, KDE4 would require at least a 2 year period of binary compatibility. I’m very confident that they’re motivated to achieve this and the framework itself is already used in JuK and amaroK.

The last talk before lunch was about NMM. It’s a research project trying to become and OpenSource project… it’s in my opinion a very difficult task. But, from what I’ve seen I wish they’ll succeed. It’s truly amazing! They made some demos… the first ones were simple, playing some audio and video files. Then they showcased the NMM backend they’re working on for amaroK… which can play video files out of the box (this feature is not currently supported by the official amaroK backends). And finally, they showed us how they can start playing a DVD on a laptop, then grab an old PDA and ask it to hand over the laptop to play the DVD too. Everything went smoothly, the video was automagically downscaled for the PDA and you could control the playback from the PDA or the laptop. Really amazing!

This last framework is really my preferred one. It’s maybe a little more complex than GStreamer or MAS but so much more powerful and versatile. Some people would surely advocate that’s overkill for a desktop computer… and they’re right of course, but somewhat miss the point in my humble opinion. You must look at the whole picture and it’s not simply desktop computing… we have more and more devices able to play media files. This framework can really break the frontiers and is a little step toward pervasive computing (also called ubiquitous computing).

After lunch, I attended the the Lars Knoll’s talk about Unicode in Qt4. It was quite interesting, especially because it showed some problem you can have as soon as you deal with several kind of languages.

I attended to Brad Hards talk. I was a bit sad to see him disappointed after this talk. He said in his blog entry that his skills for talk are not good… I believe that he is wrong on this point. The problem actually was an unfortunate scheduling. This talk was clearly made for people starting to work on KDE development… it was not the case of the audience. This talk would really have its place during the userconf… and I’d really like to see it rescheduled there.

So Brad, if you read this, please don’t be disappointed! It was really a hard task and as I said it was primarily a schedule issue. Maybe add some more fancy stuff in your slides and see if you can be rescheduled during the userconf, I’m pretty sure it would work much better!

The conference day ended with two very Qt centric talks. The first about the Model/View framework in Qt4 which seems to be quite nice and flexible. And fortunately, they didn’t fall in what I’d call the “swing trap” to use the MVC pattern for every widget under the sun. They applied this to the widgets only needing it list views, tree views and table views. The second one presented some trick and trips to go even further with Qt. It was interesting to see for example how powerful event filtering can be when mastered correctly.

In order to prepare the imminent hackfest, I finally managed to have root access on the laptop! So I had the great pleasure to recompile kdelibs and kdebase using icecream… it’s really fun and fast. I guess I’m now ready, but I really need to sleep first, I’m tired… and planning to get up late tomorrow.

I woke up early today… very tired… I admit that I didn’t sleep well last night. The Kaesespaetzle was quite good, but managed to make me feel sick during the night.

That’s why I was very very slow… I met Aaron Seigo upstairs and I bet that he wondered who was this guy not even able to understand a simple english sentence (I finally understood on the second or third try). The breakfast was quite nice and in the same fashion : slow and lazy.

Then it was time to walk to aKademy for the first talk by Eirik Chambe-Eng one of the two Trolltech founders. It was really nice and informative. Last but not least there was a lot of humour in his talk which is always a very good thing.

The second talk I attended was about the areas of KDE which would need the use of meta-data and research centric UI. The talker was Scott Wheeler who unfortunately experienced a lot of technical issues while presenting. But, he was good enough to keep the talk very informative…

Then I attended to a talk by Gustavo Boiko and Helio Chissini de Castro. They propose a way to share some code snippets inside the project to avoid duplication even more. It sounds like a good idea indeed… but it quite difficult to do it right in my opinion. This idea surely needs to be a little more refined but can turn into something really nice.

I unfortunately missed the Harri Porten’s talk I planned to attend… I was a bit late lunching and chasing for a lent laptop to be able to work since I don’t have one.

I’ve then made a hard choice and attended to Matthias Ettrich’s talk… but I’d like to attend Kurt Pfeifle’s talk about NX.

Matthias Ettrich’s talk about APIs designing was really good! This guy is really clever and master his subject… There was so many things to say that he was not able to present everything. He’s planning to write an article on the subject, and I’m really impatient to be able to read this!

Then I missed Tom Chance’s talk about the Quality Teams because I was gently discussing with Eva Bruscheifer, Peter Rockai, and two person working on the Skolelinux project : Bart Cornelis and Kurt Gramlich. We’ll see them again during the coding marathon for the KDE Debian Integration meeting.

The last talk for today was about freedesktop.org, Daniel Stone inherited this task… which resulted in some harsh considerations. I have to admit that I’m sometimes worried when I see how the things evolves, but the goal they support is quite valuable and more KDE participation would a good way to improve things. It’s just like any other project after all… its global behavior is determined by the people working on it.

And good surprise (almost since it was mentionned on the schedule), we had a “social event” with free food and free drinks (free as in free beer, not free as in freedom :p). It was really good food, and I discussed with some people.

Finally, I managed to make the laptop work correctly for my use… it still needs some tweaking but I guess it’ll be okay. The only problem is that I have to fight with this weird qwertz keyboard! It’s really unusual for me.

This time I’m really far far away from home… I’ve never been so far from Toulouse in fact.

I woke up early to take my plane… I waited for the flight to Stuttgart in Paris. And then encountered the “german way to make airports”. It’s amazing! It’s the first time I see an airport so well organized. You have nice indications everywhere, and the bagage claims are not the mess we have in France.

And then… I met my first KDE developer in real life! Peter Simonsson (psn)! He’s quite friendly and was easy to find in the airport being the tallest person there. Our adventure begun with the schnellbahn (subway)… but we managed to arrive Ludwigsburg in one piece. From there the aKademy was really easy to find. We got there to directly find Peter Rockai (mornfall) with is camera (I bet he sleeps with it :p).

Mornfall guided us to the youth hostel where we were finally able to let our luggage. We met there Joseph Wenninger and Joachim Eibl who accompanied our walking to go back at aKademy… Everything went smooth except when I managed to loose everybody… but wait it was a good thing we’ve seen a really pretty place (please ask mornfall for pictures ;p). But from what I’ve seen from Ludwigsburg I’d say it’s a really pretty city.

During the afternoon we idled trying to socialize a little except for the “lucky” people carrying laptops who were already able to hack… Please note that I was very glad to meet Kurt Pfeifle for the first time!

Finally I took some time for dinner… I tried a “typical” dishe called Kaesespaetzle, it’s really not bad…

Wow, all of this happened in one day… and the conference have not even started yet. The next days promise to be really interesting!