MAFW and Canola (and other Maemo Summit experiences)
Last week I was pretty busy with various things and I did not have much time to post about Maemo Summit, but here I go:
- The event itself was nice and well organized, probably one of the best events of this kind I have attended in the last years. Thumbs up for those who worked on making it possible. Maybe the N800 and N770 rooms were too small, but other than that it was great.
- Thanks to Nokia for letting us try the N900 before its launch date. So far I am very happy with the experience and I think it represents an impressive step forward from the N810. I’ll make sure to give you my feedback here and also in bugzilla.
I also had a presentation about MAFW there which I hope helped some people with getting started on writing plugins for MAFW. I have to admit that I had thought of giving a very different type of presentation, more based on actual source code than slides, but anyway… I just provided insight on some key ideas around MAFW and its plugins, and then I left the source code examples (which you can grab from previous post) for those willing to get their hands dirty with the details.
The session was nice, I got some interesting questions/feedback during and after the session and I hope all that feedback translates into a better MAFW at some point in the future. Some highlights (in no particular order):
- We need the Python bindings. Andrea Grandi briefed people about this on his lightning talk. There are some parts of MAFW that are not using GObject right now and that is a source of trouble for the PyMaemo team, they could use some help from people with C and/or Python experience to give the bindings a boost, so if you are reading this and think you can help them I am sure they will be pleased to know about you :). Also, I wonder if patching MAFW to be fully GObject based is still an option in Fremantle…
- Canola wants to be more integrated with Fremantle and for that they are considering using tools like Tracker and MAFW. They could get rid of the indexer tool (Tracker would handle that) and they could get access to local content, upnp media bookmarks and radio directories for free using MAFW, they could even use MAFW’s GStreamer renderer. Also, if we get new plugins for MAFW in the future these would be available for Canola too, etc.
- Talking about Canola, I was not able to attend their session because it was scheduled in parallel with mine about MAFW (this is probably my only complaint about the event), but Iván Frade (Tracker) did a good job representing MAFW there, so thanks Iván
- Since we are lacking the Python bindings, someone suggested that we should document our D-Bus APIs. That was indeed a good point, although I am not sure if all the APIs can be used directly that easily (there is some data serialization associated to the APIs that gather metadata information).
- It was also interesting to see how people would like to have a bit more of control/integration with the Media Player. That’s out of the scope of MAFW itself, but I think it is an interesting point for Nokia. I guess the idea here would be to allow third party developers to write plugins for the Media Player itself, that would enable them to modify/extend the way the Media Player works. One example of this would be a ratings plugin that would show a dialog to rate a movie when it has finished (just an example). I think this could be somehow achieved using a separate MAFW-based application, but it would be easier and nicer if the Media Player itself provides some interface for developers to do things like this.
I also had other interesting corridor chats with people about how one would write a Youtube source plugin for MAFW, how to integrate a Mozilla Flash renderer with Media Player, enabling applications to access additional metadata using Internet services, etc.
So overall I had a great time there and I think it was well worth my time. I am already looking forward to the next edition