Thursday, 5 January 2012

Open vs Closed? An Explosion of Generativity

In a previous posting I have mentioned Jonathan Zittrain's book The Future of the Internet and How to Stop It, in which he argues that the Internet needs to be open as a "generative system" to allow unanticipated change to emerge through unfiltered contribution from broad and varied audiences. His argument is that the Internet (i.e. Web) innovation needs to be open to all comers, in the same way that PC development has been unrestricted and open. No-one controls what you can do with a PC, what programs you should be able to write, to run or what information you should be allowed to process. The very processes that could control the Internet to make it a "safer" place (with regards to kiddie porn, piracy, cyber bullying, identity theft &ct.) will also tend to restrict technological development and make the future of the Internet a much poorer place - both in terms of the user experience and in terms of the future economic activity that could be developed.

In developing this argument, Zittrain and others have tended to contrast PC development (open to all) with iPhone development (closed and controlled by Apple). The first edition of his book was written before the iPhone API was released and the remarkably successful App Store(TM) was released. Subsequent editions/additions to the book have finessed the argument but by and large people still believe that a manufacturer controlled smartphone with software development policed by the manufacturer is a bad thing for innovation and hence generatively.

Historical PC/Windows Package vs iOS Package Development per year
Is this "received wisdom" supported by the evidence? The chart to the right compares the annual contribution of software developers on the Windows PC platform available from (a major software portal since the early days of the Web) and iOS iPhone/iPod/iPad platform available from Apple's app store, and apparently shows an order of magnitude more development being supported by the closed environment.

Now PC software is available from thousands of sources, not just this single aggregator, and so the number of Windows packages here is clearly underestimated, while the iOS figure is accurate (by the nature of a closed, single manufacturer environment). Still, it is not the number of downloads which is important, and which scales with the number of distribution channels, but the number of software packages that have been created. Since is such a significant source of PC software, we might expect that it would provide a not-insignificant fraction of software that is available to the general public.

So, given the arguments made about innovation and open platforms, it is interesting that there is such a difference between these figures for the two platforms in favour of the closed environment. That might suggest the amount of innovation stimulated by the iPhone is significant in comparison to the PC, that the development of the next generation of Web environments could be triggered by an iPhone-like ecosystem and not throttled by it, and that the future of the Internet is not so alarmingly threatened as some have thought.

This naive investigation and its results are an excuse for further investigation into how we theorise and predict the emergence of future web developments. The Web, after all, is not defined by the particular experience of a browser on a computer (desktop, laptop, netbook or smartphone), but by the interaction of informational and social agents.