RIM: APIs are crucial, enterprises are the target
Blogging from the Open Mobile Summit, we were interested by some of RIM Senior Vice President Alan Brenner's remarks about developing for mobile devices.
Brenner argues that the fundamental use cases for mobile applications are very different to those for desktop apps; whereas most desktop applications are "sovereign", demanding your full attention as you analyse data, process e-mail, edit documents, work with graphics, play games, write code, etc, mobile applications are usually transitory. You're doing something else when you briefly stop to check e-mail, send a text message, or look up information from the Web.
Secondly, mobile applications are usually an adjunct to some network-based resource - rather than doing the processing themselves, they send or receive data to and from a remote machine and provide the user interface. This only makes sense, given the form factor, intermittent connectivity, and battery restrictions - which will of course never go away, being based on physics and human factors. After all, a telephone is meaningless without a softswitch to talk to; computer deployment caused computer networking, but the process was reversed in mobile.
This is where Telco 2.0 comes in; if mobile applications are fated to be dominated by a mash-up paradigm, driven by the fundamental constraints of the medium, it's going to be crucially important to provide the richest possible set of APIs for them to consume. Mobile applications are increasingly intended to lash together several Web services and wrap them in a mobile-expedient user interface. If Brenner is right, they'll always be like that; systems like Palm's WebOS, which makes the whole user interface a Web browser interacting with both local HTML objects and remote Web sites, will only make this more so, as the user experience blurs between local and remote elements.
He's also concerned that a large percentage of "thousands of tiny applications" on the various app stores are used once and never again, which must surely have implications for the possible user value over the long term. Thousands of small applications isn't a problem; that's essentially how all UNIX systems work, and they are the great survivors of IT. But a lack of real user value is.
This suggests to us that value exists primarily when highly specific problems are solved by adapting general-purpose tools - like the famous quick-reconfiguring machine tools of Toyota. Put it another way, those tiny applications will be used and paid for in one particular field - the enterprise. Everything is a communications-enabled business process.
But one thing hardly anyone has discussed all morning is voice, the original and best of CEBPs.