Web applications may fracture the web

The web is an application platform like no other. It is available worldwide and can be accessed from an incredible range of devices using open standard protocols. It is not owned or promoted by any single company or government, yet after little more than a decade it has amassed over five hundred million users.

However, many of those who create web applications are not content with the web as it exists today. They see the web browser as being overly simplistic, a hindrance to the development of ‘rich’ web applications that can effectively compete with native applications. Some common grievances include:

The restrictions imposed by HTML forms and the need to implement every scrap of form logic in JavaScript.
The inability of web applications to manipulate multiple windows and dialog boxes in the same manner as native applications.
The primitive graphical control offered by HTML and CSS, which lack support for vector graphics and common image operations such as blending and filters.

The recent W3C workshop on web applications gave technology companies a chance to address these grievances and present their visions for the future of the web. However, the resulting lack of consensus in the visions that were put forward raises the possibility that the web may fracture into several disjoint application platforms.

The Triple Alliance

The Opera-Mozilla-Apple triple alliance of alternative browser manufacturers has been carrying the flag of web standards support forwards together for some time. Opera, Mozilla and Safari have advanced as one, particularly in their ongoing support for CSS, which has culminated in CSS 2.1 finally reaching Candidate Recommendation status this February.

Opera and Mozilla submitted a joint position paper to the W3C workshop on web applications, proposing that the W3C build on its success with XHTML, CSS and the JavaScript DOM. They stressed the value of backwards compatibility with Internet Explorer, and suggested that it would be prudent to extend XHTML in ways that could at least be emulated in the browser that has more than 90% marketshare. Apparently though, this did not strike a chord with the other W3C members.

Following the luke-warm response, the triple alliance has been formalised publically with the announcement of a new web applications working group that is not part of the W3C. The Web Hypertext Application Technology Working Group (WHAT-WG) will develop specifications that will then be submitted elsewhere for standardisation.

While it may seem bizarre to have the manufacturers of actively maintained web browsers forced to conduct their work outside of the W3C, in the long run this may be for the best. They are a small group with a determination to extend the web by evolution not revolution, an attitude no longer found in many of the more bloated W3C working groups.

With the expertise and motivation that the three browser manufacturers bring to the table, it is likely that the specifications developed by WHAT-WG will quickly and quietly become an accepted part of the infrastructure of the web. However, there may be more than one web to choose from by then.

The Evil Empire

In contrast to the incremental approach of the browser triple alliance, Microsoft hopes to fix the limitations of the web by replacing it. Longhorn, the upcoming new version of Windows, is slated to include a slew of new XML-based technologies designed to make it easier to create distributed applications using the .NET platform. Microsoft has also hinted numerous times that Internet Explorer will not be upgraded to improve its support for XHTML or CSS, nor extended to support SVG or XForms.

Microsoft will push hard for companies to adopt Longhorn. If the platform fulfills its promise of making Windows application development easier than ever before, it is certain to be deployed on corporate intranets the world over. However, it will take a long time for Longhorn to overtake Windows XP for home use, due to its demanding hardware requirements and the fact that most home users don't upgrade operating systems until they buy a new PC.

This suggests that Microsoft's new platform will be confined to applications developed for internal corporate use, rather than those intended for public consumption. It will join a long list of successful Microsoft technologies that are not widely used on the public web, including Visual Basic, COM, ActiveX and the .NET platform. But the ability of Longhorn to dominate even the corporate web may be challenged if a competitor puts forward a new platform based on open industry standards.

(To be continued).