At the Symbian smartphone show....

Trekked out to the desolate wastelands of East London, to the barn-like Excel conference centre today. Been listening to new Symbian CEO Nigel Clifford wax lyrical about the potential for extending its existing addressable market. The mobile phone OS company is intending to push down into the next tiers of devices, usually referred to as featurephones.

It is also loudly pitching mobile operators on the benefits of consolidating their supported handset software platforms. At the moment, most operators implicitly endorse a dozen or more embedded and "open" OS's on the phones they sell, largely because they leave the decision up to their handset suppliers. This, Symbian claims reasonably, drives up the cost of selecting, customising and testing phones. So, "obviously", the ideal solution is to become a Symbian-only operator, benefiting from the wonders and efficiencies of open operating systems....

But it's not that simple.

I remain distinctly unconvinced about the chances of convincing operators to become much more proscriptive about which OS's they support, despite this logic. The only operator that has really done this successfully is NTT DoCoMo in Japan, which develops its own flavours of Linux and Symbian OS and UI, and licences these back to the manufacturers of its 3G FOMA phones.

Other operators have either specified certain bits of the "application layer", (eg telling suppliers to use certain types of Java or Qualcomm BREW on top of their base platform), or have just written lengthy specification documents about how phones should behave.... and left it up to the manufacturers to choose their own way of achieving these ends.

Yes, in theory it would be great for Vodafone or O2 to force Nokia, Motorola, LG et al to just use Symbian Series 60, or Windows Mobile 5.0, or some variety of Linux. But in my view this just won't work.

Firstly, there will be certain customer segments that make their own choice - especially enterprise customers. If the IT director wants Microsoft phones, they'll get Microsoft phones. If their sales force has car kits that only work with XYZco's handsets, they'll get XYZco phones, irrespective of whether the operator doesn't like the embedded OS.

Secondly, there's the RAZR factor. No operator is going to want to pre-judge 2007's coolest phone now. And when it arrives, and their customers are clamouring for it, they are not going to say "sorry, it's got the wrong OS, you'll have to churn to our rivals if you want one".

Thirdly, all the manufacturers have got huge resources, personnel, intellectual property, customer loyalty and product roadmaps tied up in their midrange proprietary platforms, and they won't be binning this any time soon. Symbian's biggest competitor isn't Microsoft, it's the Nokia Series 40 team. And the Motorola Java-Linux team, and Samsung's in-house OS people are also pretty much entrenched.

Lastly, the cheapest phones will always be non-smart, knocked out for $25-50 by an ODM in Taiwan or China. They may well be cool & funky, and they'll appeal to pre-pay customers or those that just want a voice device. Forcing an expensive phone on a $15 per month user is ridiculous.

Overall, I can see some operators logically wanting to reduce the number of (implicit) OS suppliers supported via their portfolio of phones. But I just can't see it making it past the drivers of customer demand (ie cool devices) and inertia from their suppliers. Add in the fact that most users neither know nor care about their phones' "smartness", and the idea of Symbian taking over the world looks as far-fetched as ever, although its sales volumes will certainly rise. Of course, all these arguments are also true for Microsoft and the increasingly-fragmented mobile Linux fraternity as well. Bottom line? Full-blown smartphone OS's will remain niche, especially outside Nokia and DoCoMo.