Mobile Development: which platforms shall I support?
30 March 2011, 12:17PMNote: Many of the points below relate to development decisions taken within Palringo, although many of the discussions took place across multiple companies.
Related to the Microsoft/Nokia announcement last month, the eternal debate over mobile web vs. mobile apps and the recent discussions about RIM supporting Android apps on their new tablet device, I’ve thought through all of the various platform support decisions I’ve been involved in. Most recently, the debate is about when do I stop supporting Symbian and when do I start supporting Windows Mobile 7, but on mobile it has always been an issue, looks like it will continue to be an issue, but also seems that increasingly developers can choose to just support iPhone, Android (and maybe Windows Mobile).
Here is a recap of past decisions:
- Symbian S60/UIQ? At one point, maybe in 2006/2007, a developer would seriously have had to consider whether or not it made sense to support the UIQ variant of Symbian (the UI owned by Symbian before Nokia bought Symbian). I advised focusing on S60 first and then re-evaluating UIQ. It was tempting at one stage when UIQ was acquired by Sony Ericsson and, if I recall right, Sony Ericsson and Motorola were giving UIQ a big boost. Surely that was the time to commit and spend the money on development? No – within a few months, it was completely dead. Gone. Any money that a developer invested in catering for UIQ totally wasted.
- Meego/Symbian^3/Qt: This has been a topic of discussion for the past few months within companies I’m close to and the recent Nokia announcement justifies the dithering. I’d be angry if I’d invested substantial time and effort into building something.
- iPhone? Back in 2007, this was a big question. Before the platform was opened to developers it was clear that the iPhone had been done well and it became quickly clear that the development platform was sensible and that Apple were serious about creating a great target market for developers. Palringo made an early move to develop for iPhone and was one of the early apps available in July 2008 and the commitment paid off.
- What about Android? In June 2008 I wrote that I was sceptical about Android. To be clear (and try to justify myself!), the news in June 2008 was of a wave of Android phones “by year end” and of the big coming together of industry players to support Android. Unlike the decisive move to support iPhone, I advised a cautious approach. Whereas with iPhone, Palringo hired a MacOS developer and got him focused on iPhone, for Android they asked an external contractor to have a look and do some development. In April 2009, Palringo released a half-hearted attempt at an iPhone app built by an external developer. No one liked it, but it also felt like there was no big pressure to do a massive re-write to get things right despite the fact that the industry was buzzing with news of the massive wave of Android devices due by year-end (different year this time). By the end of 2009, though, it was feeling like Android was something that did need to be taken seriously and by mid-2010, Palringo had allocated significant development resource to get things right on Android. That process is still running and, given appropriate resources, it would certainly have been nice to start that process earlier.
- Then there is WebOS. I understand that some developers have developed for the WebOS devices and I’ve heard that it is an OK platform to develop for. It has never been seriously considered by any company that I’m connected with, although the recent announcements by HP are interesting enough to make me think it needs to be reviewed. Let’s see if HP can get devices in to the market.
- BADA: Interestingly I haven’t seen many people mention Samsung’s new platform BADA for a while. I’m not sure I’ve even heard Samsung mention BADA for a while. It has been considered and is regularly reviewed, but not widely supported.
- Windows Mobile 7: Much like the original view of Android, the approach has been a wait-and-see strategy, although in Palringo’s case this approach has been made easy by the fact that the Windows Mobile 7 platform doesn’t even provide the functionality required for an app like Palringo’s to work (give us a call when socket based connections to servers are supported).
I cannot believe how long that list is: in the past 3-4 years, a company like Palringo has had to make more than ~12 major decisions about supported platforms. Building for a new platform and supporting it is a big job with a cost of maybe £100k+ per platform for a typical app.
What about the end-of-life decisions for currently supported platforms? Here are some current issues:
- J2ME: Still widely used, especially on low end devices. Has always been a pain in terms of limited performance, limited functionality and massive fragmentation (often needs tweaking on a per-device basis). It is becoming difficult to find J2ME developers because they all want to work on Android.
- Blackberry: Has tended to be addressed with the same code base as J2ME, but requires some dedicated effort to handle BB specific capabilities and UI issues. It has never been easy to find a Blackberry developer and is becoming more and more difficult. The support for Android apps on the Playbook is an interesting move. Core Blackberry development is still something I view as worthwhile so long as there is a strategic fit, but they really do need to work on clarifying their strategy and engaging with the developer community.
- Windows Mobile 6.5: Still fairly widely used in enterprise. Very uncertain future and likely to die out entirely from a consumer perspective. Definitely one for end-of-life management.
- Symbian: After Nokia’s announcement, anyone who has invested in building a Symbian application is probably exasperated enough to walk away from Symbian and Nokia entirely. However, there is a big, existing installed base, there will be further devices out this year and second hand S60 devices often end up being recycled into developing markets where they have another couple of years’ life (Nokia says ~200m devices in the market now and expects to ship a further ~150m – that is compared to Apple approaching the 100m mark for total iPhones shipped last I counted). It isn’t worth starting work on Symbian, but might be worth maintaining an existing Symbian app for another couple of years.
Mobile isn’t easy. Thinking through that list, it is hard to blame the devs that just stick to iPhone and/or Android and forget about the rest, although there are still advantages in being able to efficiently support a wider range of devices.





