Following our Europe Open Source in mobile WIPjam tour, here’s the second blog in a short series summarizing the tour.
So if we do not want openness in mobile to equal opensourness soon can we really limit it to opensourceness and exclude other aspects of openness such as open development, open governance…
The first answer to the question came from a very unexpected angle and (un?) surprisingly from a non mobilist Tomcat from Mozilla. [For background information Mozilla Christian Sejersen was presenting on Fennec (Firefox for mobile) at the show and one of our discussion leaders at the WIPJam.]
The conversation had started very badly though on the slippery subject of certification and signing. The subject is fairly commonly referred to as the rock and the hard place of debates between mobile application developers. And it usually takes strong moderational talents to navigate around them… On that day I failed… But when the subject is usually the beginning of an endless (if partially justified) rant against operators and device manufacturers, despite their efforts to facilitate the lives of mobile developers* this time it actually revealed a gem… A comparison with the extension certification process as operated by Mozilla for the most famous Firefox extensions (before you click on the link believing that this is pointing towards a hugely interesting extension for Firefox this is actually a link to the process…)
Let’s imagine the world of mobile applications adopting the Firefox extension approach tomorrow:
- Developers write their applications, submit it to a sandbox / beta mobile app store. Obviously before doing so they are able to test them extensively with a series of tools, looking at memory leaks, compatibility test, traffic analyzers… all available in open source.
- A group of volunteers (editors) pick up the various proposals from the queue to evaluate and test them and communicate directly with the authors to discuss potential issues (with nice recommendations to editors such as “Remember that your comments go to a *real person*, so try to be friendly”. Or more and more let the public know that the apps are in beta and use crowd testing.
- Once approved by peers the application is published unsigned onto the mobile app store, all applications are displayed on the main page using a publicly disclosed mechanism.
- End users discover the apps on the appstore, a bit like they do today, by category, recommended, featured… and they trust the appstore just like they do today.. they trust that all the verification work has been done and therefore are more than happy to ignore warnings saying unsigned application… and despite that… they get a fully functional application
Simple! Happy developers! Happy customers! Open!
Obviously various mobile certification programs have started implementing some of this approach (do not hesitate to post spot the difference type comments), but still no one can claim to be this transparent and open today.
A pipedream in mobile? Completely unrealistic in the wireless world? How do me make money if nothing protects apps? Maybe, and all good subject of a lively debate here or at our next WIPJam session… One thing is sure though… something like this could come to a Fennec near you in the coming months…
* Background reading on certification :see the new Java verified process , the Symbian Signed 2009 roadmap, the Android Market, the well documented Apple process and each operator specific variety).
Tweet This Post
2 Responses to “Back from LinuxTag… with a Eureka feeling!”
Leave a Reply