6 March 2010

Fragmentation beyond good

Filed under: architecture, developer experience, fragmentation — David Wood @ 6:53 pm

Hmm, I thought I’d finished writing about mobile fragmentation, but the topic keeps rumbling on.

Some comments from distinguished industry colleagues prompt me to address this topic one more time.

My fellow Symbian co-founder Juha Christensen suggests “Fragmentation is good“:

I believe that already in 2011 we will see smartphones outsell “lesser” phones. By the end of 2012, there will be an installed base of over one billion smartphones. That year along, over 600 million smartphones will be sold worldwide.

One of the things that makes phones so different from PCs has to do with micro-segmentation. Like we’re getting used to hearing “there’s an app for everything,” we are a year or so away from “there’s a phone for every person”. Sony Ericsson’s newly released XPERIA 10 Mini, is a good examples of what’s to come. Intense segmentation of the market, opening up thousands of niches. Just like we all don’t drive the same car, the days are over where everyone in Silicon Valley either has the same BlackBerry, iPhone or Nexus One.

With over a billion people using smartphones, we will see thousands of micro-segments being serviced by thousands of different designs and usage model…

So far, I fully agree with Juha.  However, micro-segmentation doesn’t need to imply platform fragmentation.

With a good architecture, users can get lots of choice, even though there’s an underlying commonality of developer APIs.

As far as users are concerned, the platform supports multiple different interfaces – even though, as far as developers are concerned, there’s a single programming interface.

But Juha continues:

And this is where the goodness of fragmentation comes in. An operating system design comes with inherent restrictions. There is a need to make sure apps run the same way on all devices, that aspect ratios are more or less the same and a ton of other restrictions aimed a making the experience really good.

If one operating system was going to serve all form factors, all market segments, all use cases and all price points, the market would start trending towards a lowest common denominator…

This is where I disagree.  A well-designed mobile operating system can support a vast range of different kinds of devices.  There doesn’t need to be a “lowest common demoninator” effect.

There are, of course, some important benefits of competition between multiple different mobile operating systems.  But it’s a matter of degree.  If there’s too much competition or too much fragmentation, chaos ensues.

That’s why I prefer to say that the amount of fragmentation we have today, in the mobile space, is “beyond good”.

Sebastian Nyström also commented on the previous discussion, via Twitter:

Again we see the theme: fragmentation is part of a chain of cause and effect that has good results for consumers.  And, to an extent, I agree.  But only to a degree.

If the current amount of fragmentation is good, does that mean that twice as much fragmentation will be twice as good for consumers?  Or that ten times as much fragmentation will be ten times as good for consumers…?

If fragmentation is unconditionally good for consumers, should the designers of Qt (to pick, as an example, one important intermediate mobile platform) deliberately fragment it, into several different incompatible versions?

Clearly, it’s a question of degree.  But what degree is optimal?

Bruce Carney – former head of the Symbian Developer Network – raised the following points:

My $0.02 worth: don’t disagree that the industry should invest energy in standardization, but just cannot see the set of circumstances (or benevolent dictatorship) that will drive it as the value chain is too complex and fragmentation is a nice control point for too many actors…

So, I agree with Richard’s article. I have spend 10 years listening to developers bleet on about fragmentation, and if I could give them a simple message it would be “deal with it”, it allows you to find a niche and exploit it. Without fragmentation there would be a small number of winner-take-alls and most of you wouldn’t exist.

Yes, there’s good sense in telling developers to “deal with it”.  But there’s danger in that approach, too.

I’m reminded of an ongoing discussion that recurred time and again about strategy towards developers in Symbian:

  • Should we try to minimise compatibility breaks (such as between Symbian OS v8 and v9), or should we just tell developers to “deal with it”?
  • Should we try to minimise platform fragmentation, or should we just tell developers to “deal with it”?

The argument that developers should just accept things is that, after all, there was a big market awaiting them, with Symbian devices.  The pain of dealing with the inconsistencies (etc) of the Symbian world would be worth it.

However, history threw up new competitors, who had significantly simpler development systems.

And that’s a reminder, at a different level, for everyone preaching complacency about today’s mobile developer systems.  We need to remember that developers have choices.  Instead of working on mobile projects, they may well choose to work on something quite different instead.  The mobile opportunity is huge, but it’s by no means the only opportunity in town.  Those of us who want the mobile industry to thrive should, therefore, be constantly looking for ways to address the pain points and friction that mobile developers are experiencing.


  1. Interesting article, David.

    My view is that fragmentation is an elastic property – and that we have gone past breaking point for developers. The problem lies in the fact that the interested parties are not aligned in their intentions or in the value that a less fragmented market will enable.

    I defintely understand the view that the current state of fragmentation is a result of introducing too many control points. It would be refreshing if more of the major players were able to see the stifling effect this is having (speaking as someone who as run multiple developer programs, and is currently a developer). The irony is that some cooperation would yield dividends if applied properly.

    I have to say that I don’t agree with the “just deal with it” approach though – this cannot benefit anyone and will surely drive the industry to a lowest common denominator model of low innovation and poor products (iFart anyone?)

    Personally I hope that some of the major players take a good look at their positions and see where they best add value (and where they currently add most chaos). I could even see a couple of operators succumbing and willingly turning themselves into the best bitpipes they can be (and then really delivering great value and commercial rewards).

    Comment by Ed Kay — 7 March 2010 @ 10:57 am

  2. All innovation, most oftenly referred to as “differentiation”, is fragmentation. Depending on the maturity of the whole system, it’s either positive or negative. At one point in time having the fastest 3D hardware was a positive form of differentiation, even though it fragmented the developer APIs. When the system matured, the APIs standardised and the competition kept differentiating in some other way (this is a dim memory of the PC 3D graphics card wars in the 90s). In the mobile space today I doubt anything below the UI in the software stack is a consumer purchase decision differentiator – not even multitasking seems to be of great interest 😉

    I’m quite sure it’s self correcting, in an market-evolving way. Differentiate (innovate/fragment) too much and you won’t succeed.

    (The current shift in mobile from in-house operating systems to mobile versions of computer operating systems can be seen as de-differentiating. Hmm. Maybe I should make this post into a blog entry)

    Comment by Troed Sångberg — 8 March 2010 @ 4:55 am

  3. […] thing that baffled me for some time, was companies approaches to platforms. Google seems to be pursing two technologies, Android and ChromeOS similarly Nokia are doing the […]

    Pingback by The mobile data monopoly : what holds smartphones back from the masses. « Simon Kenyon Shepard :: justLikeThat. — 19 March 2010 @ 6:36 pm

RSS feed for comments on this post. TrackBack URI

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Blog at WordPress.com.

%d bloggers like this: