« New Year - New Moleskine | Main | Just to Clarify ... »
January 06, 2006
Mobile Device Fragmentation
I've had several discussions lately with different colleagues and customers about device fragmentation. I am always suprised at how many bad assumptions and generalizations are made about mobile devices. People making vast erroneous generalizations about devices is something that I've run into since we first started sending monophonic ringtones via SMS in early 2001 (Not all Nokia devices are created equal). It suprises me though that as far as mobile content and applications have come since, how few people in the industry have a firm grasp on the realities of device fragmentation. Even more suprising to me is the lack of effort to address the issue by a lot of mobile companies and professionals; and the effect that it is having on their products and their end-user experience.
Since device fragmentation is the core problem that we set out to challenge at Mobile Research, and I have some hard data to shed some light on the realities of what carriers have what and what phones are and are not just like the others, I thought I'd write up a few posts and clear up a few misconceptions about the state of device fragmentation in the US market.
I'm going to try to tackle one issue: Java, SMS, Content, etc. every couple of days for the next week, as well as throwing out some current device stats, so stay tuned.
Posted by David Adams at January 6, 2006 08:56 PM
Comments
I assume you mean device OS fragmentation?
With no sufficiently powerful forcing function pushing them towards cross compatability, hardware vendors have more or less allowed themselves to fragment the runtime environments on their devices because of a lack of self-disclipine. Picking a set of standards, whether APIs, DRM, or content formats, and sticking with them when facing customer (operator) and market pressure is tough. There are constant temptations to modify implementations just a little here and a little there in order to effect new functionality, esp. in the wireless market over the past 4-5 years.
Even within a relatively closed, stable OS like Windows Mobile, OEMs push for opportunities to differentiate, leading towards a degree of OS fragmentation. To expect compat. across the various implementations of Java, Linux, Symbian, WM, etc. is simply not realistic unless an sufficiently powerful interested party elects to turn up the heat.
Until operators and OEMs begin to fully internatize the benefits and necessary constraints of a platform approach, and think about mobile devices as hardware platforms upon which the primary value is created by software, there's no real danger that the need for your services will diminish.
Posted by: jason at January 9, 2006 12:53 AM
"Device Fragmentation" is not my term. Personally I'm not real fond of it; it's not a very good description of the problem. It has stuck in the industry though, so I'm stuck using it.
I agree totally. I don't forsee this problem going away anytime soon. Even if all the devices did was make phone calls and run Java applications, there would still be enough different implementations, bug, API's, and features to create a huge problem. As it is, the problem is compounded across markets, DRM, downloadable content, WAP/Web, Java, Brew, Music, Video, Bluetooth, Hardware, OS, Networks, SMS, MMS, etc. etc.
Posted by: David Adams at January 9, 2006 09:28 AM
Well, device fragmentation is problem in mobile development. The best solution for keeping all code in one file seems to be preprocessor. Only parts of source code that meet conditions are used for compilation. NetBeans Mobility Pack uses this approach. There exist others - j2me polish, antenna etc..
Posted by: Lukas at January 22, 2006 02:11 PM