There are five mobile platforms worth naming and the comparisons all run on the same axes. Symbian has the share. BlackBerry has the mail. Windows Mobile has the enterprise integration. Apple has the browser and the interface. Something is rumoured at Google and nobody outside it knows what.
None of those axes decides anything. The question that decides it is whether a person who writes software can put it in front of a customer and get paid.
Right now no platform can answer that.
Symbian has the largest installed base by a wide margin and the worst answer of the lot. To have your application do anything interesting on a modern handset it needs to be signed, which means submitting it, paying for it, and waiting, and doing that again for each revision. Then you have to get it onto phones, which means dealing with operators individually, each with a portal and a revenue split and a schedule. The share number is enormous and almost none of it is reachable by an independent developer with a good idea.
Windows Mobile is genuinely open in comparison. You can build something and install it and nobody stops you. There is still no shop, so getting it to a stranger means running your own site, taking your own card payments, and explaining to a non-technical customer how to install a cab file over a cable. That is a distribution problem disguised as a technical freedom.
Apple has spent the year insisting web applications were sufficient, which persuaded nobody, and has now said a proper development kit arrives in February. That is a reversal and worth noting as one. It also leaves four months in which the most interesting device on the market cannot be developed for at all.
What everyone is circling without saying is that the missing piece is a shop. One place a customer already trusts, with their card details already stored, where installing something takes a tap and the developer gets paid without negotiating with a carrier. Nobody has built it. The technical work is unremarkable. The hard part is being able to force the terms on operators who currently take the revenue and consider that arrangement permanent.
Which connects to the other thing holding this back. I wrote in May that the constraint on the mobile internet is pricing rather than speed, and the same operators sit on both problems. They control what data costs and they control what reaches the handset, and they have every reason to keep both.
So the platform that wins will not be the one with the best kernel or the largest share. It will be whichever one is carried by a company with enough leverage to take distribution away from the networks.
On that measure the ranking looks very different, and Symbian’s sixty-odd percent is worth a great deal less than it appears.