FOM: Future of Mobile

futureofmobile.gifToday, I went to the Future of Mobile conference in London. There were presentations by Brian Fling (Blue Flavor), Tony Fish, Luca Passani (AdMob), Andrea Trasatti (dotMobi), Steve Page (Mobile Commerce), Guillaume Peersman (Dialogue), Prashant Agarwal (Refresh Mobile), Daniel Appelquist (Vodafone), Charles McCathieNevile (Opera), Dave Burke (Google), Simon Rockman (Sony Ericsson), Julie Strawson (Monotype Imaging) and Matt Millar (Adobe).

Here’s what I took away that I didn’t already know…

  • There will be a new WUFURL Public Repository within a few weeks. There will be sub-sections for Java ME, Flash Lite and Video capabilities.
  • There will be a new WALL with an object model, CSS support and Java API. There will be 3 categories of markup for XHTL-Advanced, XHTML-Simple and WML.
  • Only 5% of mobile devices accessing the Internet are now WML only.
  • The W3C Mobile Web Initiative is also creating a device database and (eventually) an API. The aim is to provide more trust and confidence for companies.
  • UK mobile users currently perform an estimated 25 Million searches a month. 13 million of these are via Google and 12 million via Yahoo. Operators originate > 70% of searches.
  • Some content companies are moving off portal as operators move to advert rather than subscription based content funding.
  • Google page rank and algorithms don’t work well for mobile sites as they are less inter-linked than non-mobile sites. Hence, Google mobile search results aren’t that good.
  • Mippin is (another) example of a (Java) application that has moved to a web implementation due to problems with application fragmentation, download and to deliver via a medium people immediately understand.
  • There are some very difficult data protection problems related to network operator supply of contextual information (essential for filtering services such as search) such as place, time, age and gender. Some of these may need consumer opt-in.
  • Getting listed within the Google mobile listings is like being "in the woods with a very dim torch".
  • When phone OEMs produce phones, the initial design cost can become negligible when volumes are high. Hence, it’s often more cost effective to produce new models, especially when newer component costs are less.

I was particularly interested in the Google Android presentation as this area impacts me the most. Here’s what I learnt…

  • The phone stack will include Google apps but there’s no reason why they couldn’t be removed (it will be open source source after all).
  • There will be sync capabilities with Google web properties
  • It will be possible to run a conventional Java ME JVM to allow running of "legacy Java ME" applications. I found the word "legacy" amusing. I assume Esmertec will license the JVM to handset providers.
  • The same tools (compilers etc) are used for emulation as for the device. While I already knew the emulator was running real ARM binaries I hadn’t thought about the implications of this. As well as the common tools, I think this means the emulator has a very high fidelity (i.e. Runs the same code and hence has very similar functionality to the real thing).
  • The UI is "still in development" and was shipped as is to get the SDK out early. I think this explains some strange things in the UI that certainly won’t be readable on a real mobile phone.
  • The security model in (shipped) phones will be defined as (yet to be?) agreed by the network operator partners in the alliance.
  • Some companies in the alliance are not shipping open source versions of their components until the platform is "successful" so as to protect their IP.

I have lots more thoughts on Android that I am saving for another post – and for when I have time to write it!

With the exception of the Android presentation, I was disappointed with the heavy emphasis on web as opposed to Java/native (or even script) development. Maybe I am behind the times and the future of mobile really is the web browser. Maybe it’s something to do with the fact that many other Carsonified events are web related. Maybe it’s because the speaker list was drawn up Brian Fling who had more contacts in the mobile web arena. Whatever the reason, I believe there’s significant scope for innovation within the phone software itself and the application layer.