Communications - February 2009

Universal. By delivering presentation and user interface to a Web browser, Web applications were more universally available than were their earlier counterparts; coding to the Web platform—or using HTML, JavaScript, and CSS²¹—enabled developers to create user interfaces that could be consistently accessed from a variety of platforms;

Zero install. With user-interface enhancements delivered via the network, users did not need to install Web applications; and

Web APIs. Web applications exposed simple URL-based APIs that evolved bottom-up that were easy to develop, document, and learn and quickly became a key enabler for Web 2.0.

user-Centric access

As increasing amounts of information was moved onto the Web in the late 1990s, end users had a problem: To access all the information required for a given task, they needed to connect to myriad Web sites. This was true on the public Internet, as well as on corporate intranets. Moreover, the information being accessed had to be customized for the user’s context (such as desktop or mobile access). The desire to deliver user-centric information access led to the binding of mobile user interfaces to Web content, another example of specialized browsing. ²⁹^{, 33}

A user interface designed for a large display is inappropriate for viewing on small-screen devices like cellphones and PDAs. The distributed nature of Web applications—and consequent separation of the user interface— enabled Web developers to bind specialized mobile user interfaces.

At the same time, the need to provide a single point of access to oft-used information led to portal sites that aggregated all the information onto a single Web page. In this context, the various items of information can be viewed as lightweight Web components. The environment in which these components are hosted (such as the software that generates and manages the Web page) can be viewed as a Web container. Thus, common actions (such as signing in) were refactored to be shared among the various Web applications hosted by the Web container, a piece of software managing the

aggregations,
projections,
and mashups
are all a direct
consequence
of the user’s
need to consume
information in
a form that is
most suited to
a given task.

user’s browsing context.

Web components hosted in this manner relied on the underlying Web APIs discussed earlier to retrieve and display content on behalf of the user. But as long as such aggregations were served from portal sites, users still needed to explicitly launch a Web browser in order to access their information. This turned out to be an inconvenience for frequently viewed information, motivating the move by Web developers toward Web gadgets, small pieces of Web-driven software that materialize on the user’s desktop outside the Web browser. Such Web aggregation has moved over time from the server to the client where it materializes as widgets or gadgets.

Viewed this way, Web gadgets are specialized browsers. Rather than requiring the user to explicitly navigate to a Web site and drill through its various user-interface layers before arriving at the target screen, these gadgets automate away a large part of the user actions by directly embedding the final result into the user’s Web environment. Finally, Web gadgets have escaped the confines of the Web browser to materialize directly on the user’s desktop. Users no longer had to explicitly launch a Web browser to access the gadgets. However, the gadgets themselves continue to be built out of Web components. As an example, a typical iGoogle gadget consists of several components:

XML. A small XML file encapsulating metadata about the gadget;

HTML. The markup used to render the gadget;

CSS. Style rules to specify the final visual presentation; and

Java Script. Java Script functions used to retrieve and inject the relevant information into the HTML DOM before it is presented to the user.

Web gadgets relying on lightweight Web APIs, Rich Site Summaries (RSS) (letters also sometimes used to mean Really Simple Syndication), and Atom feeds helped the move toward spe-

26

cialized browsing; retrieving information from a Web site did not always require a live human to directly interact with the user interface. Today, RSS and Atom feeds form the underpinnings of Web APIs for content retrieval. In the simplest cases, they enable