vendredi 30 octobre 2009
As I am not really involved in Mozilla, I don't really know how it is to be part of this big community. It actually feels pretty good. During two days, I met a lot of people, had a lot of interesting talks every time (and some beers). I've met a lot of French contributors, shared my room with Nadir Kadem from Dailymotion, met Robert Nyman (really nice person), Vladimir Vukićević lead developer of Firefox, Honza Bambas who has implemented localStorage and many others.
No one looked at me strangely when I said I was contributing to WebKit (except Tristan Nitot but I'm used to it).
And strangely enough, I might have found there why I was contributing to WebKit. gerv said that "When the community feels smaller, it grows bigger". That's what I've found cool with WebKit : you don't need to know the project for a long time to rapidly be aware of everything going on in this community.
Mike Beltzner and the future of Firefox
I really enjoyed his talk and talking with him later. He seems very smart, reactive and open to many suggestions, how to fix problems. Be careful, he does speak French!
I'm both happy and sceptical about the pace acceleration of the releases. For web developers, it's really great cause you get new stuff and bug fixes faster. But at the same time, it can be painful if Mozilla can't migrate its user base at a fast pace too. Web developers don't want three or four versions of Firefox out there. By the way, I wonder if they will change their policy concerning older version support. Right now, I believe they support the current and the previous versions only.
They were discussions on the way to ship Firefox 3.6. If it's not yet set in stone, here is my proposal: ship Firefox 3.6 as a major update to 3.5.x and 3.6.1 as a minor update to 3.5.x. Why? We need a lot of users to move to the latest version and the more effective way to do this is use a minor update. But you don't want to hurt your users so you only use minor updates once you got the feedback of the first release (and fixed bugs). I know it's not a perfect solution (user is not completely in control) but it's the best I could think of.
HTML round table
I'm not gonna talk about the whole round table here but only a specific point around education. In order to keep the web moving and catching up with the proprietary alternatives, you not only need to implement new specs, you also need good information about them.
- Good documentation : I think DevMo is doing pretty well in this area.
- Good cross browser data : There's a lot of compatibility tables but none of them are up-to-date, complete and accurate. I think it should be the standards bodies work (W3C or ECMA). They edit the specs, the specs requires tests so they only need a good presentation of the results with a bit of high level explanation.
- Good examples : Showcasing the novelties is important and hacks.mozilla.org is young but effective in this area.
- Good tutorials : Showcasing for Firefox is cool but it often forgets the real world where web developers create for multiple browsers. You need tutorials. I think Mozilla should create a place to gather such tutorials. Opera has Dev Opera for that. I really like articles like Cross-Browser Inline-Block.
Firebug and Jan Odvarko (Honza)
They got a lot of cool stuff coming in Firebug 1.5. I especially like the capacity to break on many events (HTML change, CSS change, JS property change). Makes it very easy to understand how a piece of software works just by saying "hey, this part of the page changed, show me the code that did it". Very smart. The HTTP Archive format is also exciting and I hope it will be integrated in the Web Inspector (even better if I find time to do it). Don't miss the slides (and detailed follow ups on his blog)
The labs projects are very interesting. I'm still not convinced by the migration of apps to the browser (I prefer native apps using the network) but if it has to be, Prism is a good move.
The synchronization part of Weave already looks good and I like the open way Mozilla is using for all his server side efforts: develop a server, provide the protocol but also host a version. I also have big expectations on the sharing part of Weave. Something like Opera Unite (share data with friends without uploading a thing) but that can also work when you're offline (cause it's served by the Weave server which has synchronized in the background).
Jetpack is a good reaction from Mozilla but it has to happen fast. Chrome extensions will be really easy to create and ship so it will be a competitor. Not for the end user but for web developers that will try Chrome to develop extensions, maybe use it and then recommend it. I'd say don't listen to add-on developers scared by the lack of functionality, disparition of XPCOM, etc. Add a lot of APIs, get it in Firefox as soon as possible and market it as the easiest way to create extensions. You'll figure out later how to deal with existing extensions.
Irina Sandu, William Quivigier and Paul Rouget
Thank you for the organization, everything was perfect even when I missed my plane. I understand how hard it might be and being so nice while having so many stuff to take care of is amazing (I'm sure FuzzyFox will agree)