Thursday, July 26, 2007

Statistics For Morons

"According to figures announced today by the Mobile Data Association (MDA), an overall total of 45.6 million unique users were recorded as having used their phones for downloads and browsing the mobile internet in the UK throughout October, November and December last year, an average of 15 million each month."

So by that logic, after 4 months at this rate we'd have had 60m unique users (every person in the UK), and by the end of the year 180m unique users in the UK - or the UK population plus all the Polish plumbers, and the whole of the rest of their compatriates back home plus the population of Germany for good measure, all deciding to come to the UK to use the mobile web. If we take into account increasing demand - by extrapolating a straight line of course - the UK will have more than the entire world's population of unique users within a few years...

Bet Not Over Yet

A quick thankyou to Mr Mace for pointing out that my hundred quid bet is still not decided yet, nicely touching on a number of the reasons why it's actually very difficult to work out what has shipped between particular dates.

Currently I feel optimistic - this product was always going to sell to the faithful in large volumes over the opening week. It certainly won't sustain that rate so the bet all hinges on how quick the drop-off is, and whn they release v2 which should hopefully deal with a lot of the flaws (though I suspect there won't be a keyboard turning up any time soon so a lot of people will stay out in the cold).

Wednesday, July 18, 2007


O2 have just bowed to the inevitable and ditched their i-mode service - excellent news.

There's nothing particularly special about DoCoMo's Japanese i-mode service apart from the fact that loads of people use it and it makes money for everyone involved - an interesting counterpoint to the situation in the rest of the world where operators feel only they should be paid for anything and few people use their equivalent services.

Japan is years ahead of the West in many ways, but it is also very different. When i-mode was launched Japanese PC/Internet penetration was extremely low, partly for historic reasons and partly because when you live in a shoebox you can't easily squeeze a PC tower in the corner. DoCoMo presented a very culture open to new technology with a fast, simple way to access services online whilst seeding an ecosystem with inclusive open technologies - customers loved it and usage skyrocketed. With complete control over handsets - it's often hard to tell which manufacturer actually made a Japanese handset - DoCoMo ensured that all customers could find the services they wanted, and they all worked in a consistent way removing a vast swathe of development hassle. This is why i-mode is popular in Japan, despite the fact that most i-mode pages are as ugly and awkward as their wap brethren.

Almost none of these factors apply to the overseas licencees. Customers had easy access to PCs and the 'real' Internet which pushed user expectations way beyond what any phone could do, let alone phones offering what actually is a very primitive-looking service. The European markets i-mode was trying to compete in were awash with very attractive high-spec handsets, benefitting from the economies of scale that have contributed to GSM's success. DoCoMo managed to line up a handful of largely uninteresting handsets to compete, offering none of the standardisation benefits they enjoyed in their home market whilst being seriously outgunned in aesthetics and specs. The Overseas Edition i-mode wasn't even as good as the Japanese version - it was several years behind the Japanese version but also, criminally, technically behind its European contemporaries (particularly with the DoJa 1.5 flavour of Java which contains pointless restrictions). With such a selection O2 could never hope to offer only i-mode devices, therefore they had to manage i-mode (and DoJa games/content) on one side and wap (with MIDP) in parallel. Finally, users didn't really know what i-mode was or why they wanted it, so handsets were not selling because they were "i-mode" - they needed to sell on their other merits as a Trojan Horse.

Thus the i-mode experiment outside Japan was inevitably going to be a failure, but it has taken 2 years to put it down. It goes to prove that you'd be very foolish to draw too many conclusions from the Japanese market, especially if you don't actually understand all the factors involved.


Thursday, July 12, 2007


Sony-Ericsson make the best mobile UIs, Motorla some of the worst (though things may be changing, and buying Sendo was a great move).

Sony-Ericsson have been pushing in new directions with the use of the Walkman and Cybershot brands (with Wega coming soon?) for vertically targetted devices, Motorola have spammed the world's mobile shops with identikit RAZR clones (oh, and they did the ROKR).

Amid strong growth in the mobile market, Sony-Ericsson profits are up 55% this quarter, Moto is down compared to last year and may make a loss in mobile phones this year.


(OK I think it's safe to stop bashing Moto now, one day soon they'll become the plucky underdog and people will start rooting for them again - the new platform could do it if they find a nicer box to put it in)

Thursday, July 05, 2007

Open Paedo APIs? (No, We Don't Know What He's Talking About)

Ajit of Open Gardens fame has recently received some top quality smokeable materials, it would seem, as in his latest blog post he has underlined the imminent threat to the world's morals posed by - can you guess? No? Uncontrolled APIs. To quote in full: "uncontrolled access to APIs is an invitation to scammers and paedophiles." Oh yes.

Scammers I can see. If you allow an AJAX script to initiate a phone call or send a text without user intervention, you could easily have any wap page - or advert - start racking up premium charges. You definitely want to make sure the user can prevent that. I would recommend a simple system where you ask the user what they want to do, but if you don't understand how code signing works it might seem that a laborious code signing exercise like the flawed Symbian Signed could maybe succeed.

Paedophilia though? Really? It's not my expert subject, but I do know a little about APIs and I don't really see the connection. If anyone can enlighten me please do so in the comments... Maybe this requires more thought. Perhaps there is also a terrorism angle that the pragmatic mind of Mr Jaokar has missed? Perhaps he could mention it to the European Parliament next time he drops by to brief them on what is happening in the real world?

Sorry to keep bringing the pragmatic thing up, but he started it ("but I have always been pragmatic") - and it's funny to see someone who really genuinely feels strongly about being pragmatic in his approach to people (quite rightly) without being able to see the utter lack of pragmatism in his view of the technology he is associated with.

The signature comment hilights his lack of understanding - you can code sign a single binary app because it's a single binary file and after every rebuild you can pay to submit it to code signing for a few weeks and maybe get it back approved. How do you sign a dynamically generated wap/web page which references a standard script library and maybe some other embedded custom scripts, a few external scripts, and some other server generated script and data? You can't. It simply doesn't work. Suggesting it does just shows how little Ajit understands that whichi he professes to be an expert in.

Apologies to Ajit and readers for getting worked up about this, I am certain he is a really nice bloke and he does a lot to raise the profile of mobile applications etc, but I strongly believe that if someone wants to create a massive soapbox and shout from it to everyone who will listen, that person really understand what they're talking about first and has a duty to communicate clearly and correctly. Suggesting this kind of thing is trivial and essential for the market to proceed implies some due diligence has been observed, but it clearly hasn't. Rant over.


Monday, July 02, 2007

TMobile Devicating On JavaME, says Nokia?

An amusing typo on the new Forum Nokia wiki gives the standard permissions for API access on Nokia MIDP phones, and then lists T-Mobile US as among the operators who are "devicating from the standard". Which sounds awfully appropriate given T-Mobile's refusal to allow apps signed by a 3rd-party to be installed on their phones (only unsigned or operator signed - kind of kills the point in signing...) and their propensity to lock phones down and prevent users from doing nasty things like using the network from Java (operators being natural opponents of network usage...)