2009 was a bad year for me - not without some amazing highs, but the lows won out. I'm not going to write about that, as most of it isn't suitable for public consumption. Instead I'll share a quote which stuck with me through the last 12 months; it's from the Qu'ran, and it's the most wonderful expression of optimism I've ever heard.

"At the end of the world, plant a tree"

I've got a set of things I want to do in 2010. Again, not all are suitable for public consumption, but a couple of them I'd like to publish in an attempt to egg myself into doing them:

1. I want to make things again. I've not done much of it over the last year, I miss doing it and the further I get from building software, the less able I feel to influence how we do it at FP. So I'm going to spend at least a day a week working in one of our two production teams.
2. I've ended every year for the last 10 being fitter than the previous one. This year I upped my running from about 40km per month to 100-120km. I want this trend to continue. I am never going to die.

I'm rubbish at predictions; in 1993 I thought the WWW was a waste of time, and that the future belonged to gopher. Nevertheless I've managed to contribute a few thoughts on where mobile is going in the next 10 years to this presentation which Rudy de Waele put together over the new year break.

Mine are pithy cop-outs, in general - of course there'll be more bandwidth, of course more of our internet access will be mobile, and it's hard to predict the complex societal effects of technology. We didn't get it right with SMS (who'd have thought 160-character messages crowbarred into the GSM spec would alter our habits around social rendezvousing?), so our chances of getting it right with the likes of augmented reality and mass access to the internet are pretty low. Never mind, best if we all just inspect and adapt, eh?

I do think that power and battery is an underexamined topic: screens, connectivity and location technologies all need energy from somewhere, and I'm convinced that a revolution in this area will change a whole load of deep assumptions that we make as an industry over what's possible in the land of untethered mobile devices.

Go check out the presentation, there's a pile of great stuff in there. David Wood has a nice summary of it.

I love this interview with Jon French of HTC. Not because of what he reveals about their approach to the Sense UI; not because after 10 years of Nokia and Sony Ericsson loyalty, I'm a happy HTC user myself; but because of this quote:

Plenty of previously agnostic OEMs have now launched app stores. Where’s yours?
We have no plans for an app store. HTC is not in that business. For us, it’s all about how you integrate the app store – whether Windows Marketplace or Android Market – into the experience of using the device.

How refreshing; a large player in mobile with no plans to open yet another app store!

So, I've had a few days now to fiddle with with Guardian iPhone app - and I'm liking it. In an attempt to stretch my ability for unbiased appraisal to snapping point, I thought I'd do a little comparison of it to our own (unofficial) Guardian Anywhere app, for Android devices.

Incidentally, we saw a nice little uptick in usage of the latter when the iPhone app launched - and we're currently cruising at 4000 downloads (with 1700 of those being active installs). Considering that our promotion for this was a blog post here, a single tweet, and a gloriously underfunded Adwords campaign (446 clicks delivered so far), we're pretty pleased.

So, onto the comparison... looks-wise, I have to say I think the Guardian have done a much better job than us of replicating the brand, and the feel of print. The iPhone app feels like a newspaper, the Android app more like a database of stories. We've considered doing a more "papery" version for larger-screen devices like the Archos tablet, but it's unlikely to happen now - our attention is focused elsewhere for a little bit.

That said, I find the iPhone app a little "busy" - there's important navigation (search and settings) in the top corners, a navigation bar across the foot, a selector for latest and trending stories, and a feature that could be really useful (offline browsing) mixed into the foot of the main navigation. I could still find everything I wanted quickly (so my complaint feels a little churlish), but the overall sense was of extreme busyness - and I wasn't completely clear on the difference between "latest" and "trending".

Both apps make good use of the excellent pool of photography that the Guardian publish. But - and here's a key difference - from the iPhone app I can email a photo, and that's about it. On Android, I get the option to set as wallpaper or share the photo through a variety of mechanisms, highlighting one of the key advantages of the Guardian Anywhere (which is actually an advantage of Android as a platform): the interconnections between apps.

On my HTC Magic, I have apps for Facebook and Flickr installed. These apps then expose photo-sharing services to every other app on my phone - meaning that my copy of the Guardian Anywhere can automatically share photos with both of them, at zero effort for us, the developers. That's a really big win for me as a user: if I'm a member of a niche social network, my social networking app can help the Guardian Anywhere share content.

The same thing applies to the sharing of stories: I can post to Twitter easily from Guardian Anywhere, by virtue of having installed Twidroid. And I notice that within the stories themselves, links are stripped out on the iPhone (whilst being kept in on Android).

The iPhone app includes a whole load of audio content which is ace - I'm betting that more iPhone/iPod Touch owners listen to music on their devices than Android owners.

I was chuffed to see offline reading make it into the feature-set, though I can't help feeling that the value of it is reduced by the lack of running it at scheduled times in the background. One of the things I love about our Android app is that when I wake up, the news is already sitting there on my phone. With the iPhone, I have to tell the app to go grab the content now - which takes around 15 minutes. This just doesn't deliver the same sort of convenience, though lack of background processing on iPhone didn't leave the developers with any options.

I feel a mix of comfort and schadenfreude to see that downloading offline content takes about 15 minutes on the iPhone - it's the same for us on Android, and it's the number 1 complaint from our users. I'm confident that if the Guardian haven't solved this problem themselves, it's not so bad that we haven't yet...

And finally, there's distribution. I can't help but point out that the Guardian Anywhere is free, and available globally. £2.39 isn't an extremely reasonable price-point for such a high-quality iPhone app, but it's not available outside the UK, US and Canada. Whilst most of our users are UK and US, we have a sizable long tail in France, Singapore, Germany, Australia, and quite a few others, making up around 12.5% of our users.

It feels strange - a comparison of two very similar apps on Android and iPhone has ended up being a comparison of the platforms themselves, with iPhone delivering a superior overall look and feel (at a small but reasonable price) and Android making better use of background processing and connectivity between applications to improve the experience, for everyone.

Russ wrote a nice post a month or so back about the splintering of Android.

He's right, I think - we are starting to see fragmentation around Android. When new handsets or versions of the operating system launch, we find ourselves getting occasional small bug reports for our Guardian Anywhere product. Some of these bugs have been down to our not doing things in a standards-based fashion; some are down to differences in handsets (the Tattoo, for instance, has a much smaller screen than most other devices); and some... some we can't quite get to the bottom of.

And of course, it's not just the handset. In an application-rich world, every user has a different combination of apps, and Android's architecture allows them to run many apps at the same time - all of which might interact with one another, leading to difficult-to-reproduce edge cases. I'm also expecting to see issues caused by handset customisation (which operators are doing a great deal of - the T-Mobile Pulse is a nice example of this, and it's still early days)

So there is fragmentation in Android; but I was pleased last year to see Google quietly move from the naive line that "fragmentation won't happen, because it's not in the industry's interest" to "we'll introduce conformance tests for OEMs to avoid fragmentation". This seems to be paying off, so far - in that moving Android applications to new devices, even those with new screen sizes, is orders-of-magnitude less painful than moving between J2ME handsets from different manufacturers. That said, developers still need to do some work to provide graceful degradation.

Apple have done a fantastic job of presenting the iPhone as a single platform. There however, the reality is slightly different: 3 generations of hardware, 26 operating system releases (last time I counted), and a platform which in some incarnations has (or doesn't have) GPS, a loudspeaker, a microphone, or even a SIM card and therefore connectivity. Apple have done a great job of upgrading operating system versions to keep its installed base current, but even there there's some lag. The problem is at least finite, Apple being a closed ecosystem for hardware; and the lack of background apps helps developers avoid issues caused by interactions between applications.

So where does this leave us?

• Fragmentation isn't going away. Not until OEMs stop innovating and differentiating their products;
• Even wunderkinds of the tech industries like Apple and Google can't solve the problem for us. Even their kit - whether it be strictly controlled or lovingly curated - suffers from the issue, whatever you might hear to the contrary;
• As an industry we're getting better at dealing with it: platform vendors are taking more care, developers are evolving techniques and technology to cope;
• For those of us taking mobile products to market, testing across a range of devices is going to be something we continue to do as a matter of course - and we'll still be learning more than we ever wanted to know about weird handset bugs and software versions;

Many thanks to James Hugdroid, who proof-read, sanity-checked, and contributed to this post.

Update: Google have published a really nice analysis of deployed Android versions, implying that developers should be targeting 3 versions of the Android operating system.