Microsoft’s recent announcements around Windows Phone 7 Series and first details regarding application development for the platform have made me take pause to think about those of us out in the world of mobile device application development. More specifically, I’ve been thinking about things from the perspective of a developer who supports mobile device platforms in general, not just one in particular. I know I am anything but alone in this situation. An increasing number of independent developers and commercial development companies are in this situation. Additionally, the number of enterprises supporting Line of Business applications to an increasing variety of users and phones fall into this same group. For everyone involved, the dizzying array of hardware types, operating systems, development platforms and programming languages are a challenge, to say the least. That is why I am becoming increasing convinced that for many, the only real hope in being able to support all mobile device platforms is HTML 5.0.
In order to understand the true scope of the current dilemma for so many people, you simply have to break down the potential target phones from a development perspective. You get something like this…
- Windows Phone – Currently, you have native development with C++ or managed development with the .NET Compact Framework. Moving forward, you will have Silverlight or XNA. The development platform – Visual Studio (running on the Windows operating system).
- iPhone – Objective-C is the language, with XCode as the development platform (running on the Apple OSX operating system).
- RIM – A device-specific subset of the Java language, with different development platform options.
- Android – A device-specific subset of the Java language, with different development platform options.
- WebOS – A device-specific subset of HTML standards, with different development platform options.
OK – so what does this mean for a developer trying to build a single application that supports multiple phones using native development tools? Well, for starters…
- From a development computer, the best-case scenario requires an OSX-based desktop or notebook (think iMac or MacBook) with both the OSX and Windows operating systems installed. I have set up such a computer in the past, and it’s not very pretty (or inexpensive).
- Again from a best-case perspective, I need to know at least the fundamental aspects of -
- .NET (then the derivatives for the Compact Framework, Silverlight and XNA)
- Objective-C (C or C++ for starters, but a whole lot of differences)
- Java (then the SDK-specific derivatives for each device)
- HTML (and then the specifics for WebOS)
Folks – that’s one specific computer and FOUR programming languages (with a lot of additional knowledge on top of the language basics) just to get my single application to all the major phone platforms. And remember – these are “best-case” scenarios. I didn’t include those applications that are very hardware-dependent (physical buttons versus touch screens, screen resolution, etc) or graphics-intensive (using GDI versus OpenGLES versus native, etc). Of course, I left out the biggest challenge of them all regarding development – having to re-create one application many times. Even under the best of circumstances (an application where the bulk of application logic resides in a centralized location accessible via the Internet), there is all the work on the client/device of the application to perform. Even under the best of circumstances (Java to Java, for example), there is still lots of platform-specific work to do).
With all these challenges in mind, what’s a developer to do? Well, there are several options -
- Pick your platform(s) is stick to it/them. For many independent developers, this is the only viable alternative. The time and effort for a single person to develop one application for multiple devices without a guarantee of return on investment makes for a simple decision – choose the platform or platforms I can reasonably manage from a time and cost perspective. For larger development groups, there is still the challenges of identifying and allocating the proper resources (developers) to the development efforts, never mind the challenges of keeping everything in sync to assure that work on Platform X doesn’t break or supersede Platform Y. For enterprises, there’s the cost of development, hiring and training the resources required.
- Leverage multi-platform development platforms. This is primarily an enterprise-focused option that is becoming more popular. There are solutions that allow you to ‘write once, deploy everywhere” when it comes to phones. However, these platforms often come with new challenges. In addition to the sheer cost of such solutions (limiting their consideration to larger enterprises with IT budgets capable of purchase), there are infrastructure and capabilities (think “Jack of all trades, master of none” when it comes to these platforms when compared to native development) to consider.
- Develop mobile web application development. As any developer knows, the current state of mobile web application development comes with a number of limitations. First and foremost is the issue of being offline from the Internet and the “red flag” this poses for many applications. Despite all the marketing hype, guaranteed mobile device access to the Internet is still a myth (just think about all the times you’ve had dropped calls on your cell phone). Heck – guaranteed wired Internet access is a myth; just think about your home computer and your ISP. Many application rely on being able to work when no Internet connection exists. These applications not only need to function from a user interface perspective; they also need to be able to safely store enough of a cache of data locally when the the connection is lost to be able to allow the user to perform their work. For the most part, most mobile phones do not implement the capabilities within their respective web browsers to make this a reality.
The current state of mobile device web browsers also makes for a present-day situation I like to refer to as the “Microsoft/Netscape Syndrome”. Anyone who has been developing web applications can remember the days when the differences between the two most popular web browsers at the time (Internet Explorer 3.x and Netscape Navigator 3.x) were so different in what HTML standards they supported that as a developer you were required to essentially build two versions of a web application if you wanted the widest usage. NOTE: Yes, I acknowledge that this is STILL going on today in the desktop browser world – I just like to refer to the first time that this became an issue that truly affected the development world in a broader scale ;-) Today’s mobile web browsers mostly suffer this same fate. Their capabilities vary by platform, making even the simplest of application development tasks difficult and requiring a lot of conditional coding and rendering logic based upon browser detection (the term “browser sniffing” is has relevance more than a decade after it’s creation).
Now after saying all of this (including the current challenges around mobile web development), you may be wondering why I think that yet another HTML specification may provide some relief for multi-platform mobile device developers. Well, there are a couple of things around HTML 5.0 that can address these issues -
- Reducing fragmentation that has occurred in the years since HTML 4.0 was ratified. A lot of time has past in the years since HTML 4.0. During this time, a lot has changed with regards to the Internet and how we use web applications. We’ve seen fragmentation in addressing these issues as a result. Don’t get me wrong here; I’m not saying that HTML 5.0 will solve all the issues. However, we are already seeing some vendors in the mobile space betting in some way, shape or form on HTML 5.0 and moving away from their own individualized solutions to problems, instead building in HTML 5.0 support in their own mobile web browsers. Heck – Palm has practically bet everything on HTML 5.0 with development for WebOS. As I mentioned earlier, WebOS application development is grounded in HTML standards – HTML, JavaScript, CSS, etc. If that makes you wonder how you could possibly develop applications that work offline, then you need to consider…
- The HTML 5.0 Database specification. I will leave the details to this specification to the W3C. However, if you care to look at the HTML 5.0 specification, you will find a great deal of changes dealing with application caches, application sandboxing and databases. All of these changes account for offline HTML applications as well as traditional online access. In essence, the W3C is, with the HTML 5.0 specification, finally officially addressing the use of native HTML as an application language, rather than simply a connected web application language.
Palm is not alone in moving toward HTML 5.0 for mobile technology. Google (via Android’s browser) and Apple (starting with the iPhone 3.0 and Safari) have started implementing HTML 5.0 standards. One example is with the HTML 5.0 support for location. If you’d like to see this in action and have either an Android or iPhone handy, check out Microsoft Device Application Development MVP Richard Jones’ example (complete with source listing). Bottom line – if the HTML 5.0 as currently written is ratified and if the mobile device platform players implement the standard, it will be much more possible to develop cross-platform applications that meet offline requirements than ever before. While I am certain there will always be platform-specific APIs (likely exposed via JavaScript) to deal with, I think that learning one base language and device-specific extensions is certainly more manageable than four languages that exist today.
Of course, HTML 5.0 cannot solve all multi-platform development problems. There will almost always be application scenarios that require using a development model and language that is “closer” to the device. Gaming applications, for example, commonly require complex graphics processing. While HTML 5.0 is making strides even in this arena, it is likely that it will not supplant the need for more native development solutions that are optimized for and tailored to a specific device. For many other applications, however, HTML 5.0 provides the potential to enable development of applications that can meet the challenges of modern mobile applications, and all under a single development language umbrella.
I applaud those device and mobile operating system manufacturers that are embracing HTML 5.0 as “early adopters”. To those that are not, I strongly urge them to consider not embracing this standard might mean. Early on, I mentioned the “pick your platforms approach”. If HTML 5.0 is adopted by everyone but you, it is more likely that you will be the “odd man out” for a potential audience of thousands of developers. Here’s to hoping that is not the case for any one manufacturer. Here’s also to looking forward to HTML 5.0 final ratification. And finally – here’s to hoping that HTML 5.0 really can live up to addressing one of the greatest challenges posed to mobile device developers today.