IBM Position Paper on Mobile Ajax

Situation analysis

Today's mobile devices offer a wide range of Web capabilities. An over-simplified characterization is as follows:

As a result, from the perspective of a content developer, today's mobile world is highly fragmented. The Mobile Web faces numerous other challenges. The lack of unlimited data plans often precludes Web access for financial reasons. Slow data rates and high latencies can result in painful user experiences. "Walled gardens" might preclude access to the entire Web. Most Web site are designed for desktop access, which often means the Web sites will not work well or not at all on mobile devices because they exceed memory limitations, take too long to download, don't fit on a small screen, and/or depend on a mouse or full keyboard in order to operate.

On the other hand, mobile devices also offer fresh opportunities to the Web community, especially when combined with the Web 2.0 phenomenon. Mobile devices are well-suited for data capture in user-generated content scenarios because you carry them with you, are (nearly) always connected to a network, and typically support capture of most common media types (text, audio, images, and video). Often, you can determine the latitude/longitude location of the mobile device via techniques such as GPS. Furthermore, there are potential synergies with device-native features, such as address books, text messaging, instant messaging, voice messaging, and photo management.

Moore's Law and Competitive Pressures

As Moore's law dictates, with each passing month, higher percentages of mobile devices support Web browsers similar in capability to those on desktop machines. Early mover mobile devices, such as the Apple iPhone and Nokia Series 60, are educating the community about what is possible, which will drive demand and put competitive pressure on the rest of the mobile industry to offer similar Web capabilities. The browsers in the Apple and Nokia devices are from the WebKit open source project, so there are no licensing barriers impeding ubiquity. Increased demand for Web support will pressure operators to offer friendlier rate plans, improve their data speeds, and address latency issues.

The long tail of Web content

Because only a small subset of mobile devices today offer full Web support, content developers have to make choices. Most of the existing content on the Web was designed under the assumption that a desktop browser such as IE or Firefox was available. Web developers have the following choices: (a) wait until Moore's Law makes full Web browsing ubiquitous on mobile devices, (b) redesign their Web site in a device-independent manner with adaptation software that generates suitable content for different devices, or (c) develop alternate versions of their Web sites using alternative technologies (e.g. XHTML-MP).

Only a small percentage of Web developers they can justify the expense and learning curve to pursue alternatives (b) and (c) The long tail of Web content is likely to choose (a) and wait for devices to offer better browsing support.

It is important to point out that a significant percentage of the Web operates in intranets and extranets, often to serve needs of businesses who have to justify their IT expenditures. The majority of IT applications fit into the long tail. Most intranet and extranet applications won't receive funding to pursue options (b) or (c), and instead the IT managers will wait until mobile devices support desktop browsers.

Corporate developers may make movements in the direction of option (b) as they seek to takle other problems that they are faced with such as the need to satisfy Accessability requirements. Accessability is becoming the subject of legislation in an increasing number of countries and will not be optional for businesses that operate in some countries.

Even with high-end mobile browsers, there are issues to address

Mobile browser vendors are designing their own approaches to optimizing the Mobile Web experience. Different mobile browsers offer different innovations to address small screen size issues with names such as "Small-Screen Rendering", "Smart-Fit Rendering", "Mini-Map", and double-tap zooming. Sometimes the mobile browser vendors invent extensions to HTML, CSS or DOM to allow content creators to achieve optimal results. The devices themselves have different combinations of keys, buttons, touch screens, joysticks, and wheels. As a content developer, the input capabilities are not only different than desktop computers, but different across the various mobile devices. As a result, even for the long tail of the Web which chooses option (a) and wait for desktop browsers to propagate on mobile devices, there will be fragmentation and issues that need to be addressed for developers to adjust their Web sites to work suitably across a wide range of mobile devices.

Recommendation: focus on "one Web" and incremental value-add

W3C and OpenAjax Alliance should accept the inevitability of Moore's Law and recognize that consortia take years to develop specifications. Any initiatives must take into account the trajectory of the Mobile Web. Therefore, efforts should focus on the most likely long-term eventuality, which is that ultimately large numbers of mobile devices will be capable of running Web browsers similar in capability to those on desktop machines. The primary thrust should be in helping the industry learn how to create Web content and Web browsers such that end users will be able to successfully access the full Web from a variety of different mobile devices. Likely candidates for activities include: (a) incremental work on existing W3C technologies (HTML/XHTML, CSS, DOM, etc.) to address the special requirements of mobile devices, such as small-screen issues and different interaction capabilities; (b) incremental work on MWI Best Practices to take into account the trend towards Ajax-capable mobile browsers and the Rich Applications that become possible; (c) new standards work as necessary to unleash the potential of Web-connected mobile devices, such as integration of device capabilities with browser-based scripting; (d) definition of modality independent events; and (e) education and evangelism.