Surveying its conspicuous failure to make a dent in the mobile phone market or even to ruffle Nokia's feathers ever so slightly, Microsoft can console itself with one thought; as has been typical of Microsoft's battles for some years now, this has been a Microsoft invasion of somebody else's turf, so failure doesn't threaten the rent from the main ranch, and that rent can be used to regroup, gather strength, and invade again and again till you finally win. But what if the defending party was not prepared to just sit around? What if it decided it was time for a counter-attack?
That, it seems to us, is precisely what Nokia is planning. Its mobile phone home turf faces diminishing returns and lower margins as the market matures, so it has to extend its reach into related areas, turning its description of itself as a producer of connected devices from a slogan/pious hope to a reality. So enter Jon Bostrom, Nokia chief Java architect and Nokia rep on the board of the OSGi Alliance. If people have heard of OSGi at all, they quite likely think of it as a standards group for home automation and device connection, but it's more properly about interoperability and management of devices, both fixed and mobile, and that's why Bostrom is on the board, and why Nokia rejoined OSGi earlier this year. It's currently one of those 'most people except Microsoft' organisations, but we note from the delegate list for this year's World Congress that Redmond, diligent as ever, sends along its minders.
OSGi covers devices in all shapes and sizes, but the most immediate relevance from Nokia's point of view is in the business connected device market, where if anybody's in the driving seat right now it's Microsoft with PocketPC, with Palm as the challenger. Yes, we know you could put that the other way around, but even if you do you have to at least accept Redmond money eventually beating Palm to death as a possibility. Can Palm resist the Microsoft tide on its own? Doubtful. Can Nokia mount a successful invasion and reverse it? Maybe a different answer.
Bostrom, reasonably, doesn't accept that this is anything like a mature market with an established leader. Companies with an 'official' PDA policy are at the moment early adopters, and given that the real market is going to be for "mobile devices that are network connected and the IT manager doesn't need to care," you could even argue that the space is virtually unoccupied. Bostrom views PPC as simply having transferred the security problems of Windows to mobile devices, and categorises Microsoft support as woeful ("Microsoft gives you a CAB file, says go away, you take care of that"), and expects TCO studies from hell to be helpfully documenting this a couple of years hence.
Attacking this market is therefore not just about building the devices but about getting the management systems for these devices right as well. As a matter of fact, in theory it needn't be about building the the devices at all, but if the likely dominant player isn't going to want to play with your management systems, then in practice you're going to have to build the devices and beat them on volumes. Bostrom doesn't directly say that, confining himself to saying that there will be successor products to the Communicator (coincidentally...), but it's clearly necessary, and equally clearly the form factors have to be a lot more diverse than just Communicator.
We also need to think about service delivery in the same breath as management here. To some extent this is where the OSGi connection comes in, but we also have this, JSR 232. This "will define a component management framework that will allow mobile devices based on the J2ME Connected Device Configuration to evolve and adapt their capabilities by installing new components on demand. These components can be a combination of active elements with no user interaction (services), active elements with user interfaces (applications), and shared libraries (both native and Java). The framework will enforce a predictable lifecycle model for these sharable components that will encompass install, start, stop, update, and uninstall. The framework will also provide for multiple applications to coordinate the use of single-access resources such the device display. In order to ensure a safe environment, these components will be controlled via a mandatory security model based on the Java 2 Platform security mode." And note that it's a joint submission by Nokia and Motorola.
Java development for mobile devices has been successful so far on the 'sandbox' principle, where it's sufficiently generalised and platform agnostic for it to at least come close to 'write once, run anywhere.' But one of the next stages is to take it out of the sandbox and to allow it to tie in more closely with the platform itself, and with the services the platform receives. So you could maybe see this as being as much 'Java invades Pocket PC' as Nokia invades. But if it's Java, Nokia probably builds the tanks anyway.
Bostrom himself has a substantial rap sheet as a Java architect, so much so that The Register asked him if he really worked for Nokia, or if he was just a Sun hitman on secondment. He did used to work for Sun, but says he now really works for Nokia. Why? He'd done Java implementations (for example, for Sprint and DoCoMo), but coming to Nokia gives him the control he didn't have in previous iterations. It allows him to "create the architecture inside the device and finally get it right," he said. Which in our view just about clinches it. And if you need any more, we also asked if Nokia was proposing to move into the space formerly inhabited by Psion, now the latter is as CE company. He smiled, and told us we could speculate. ®