Building a better browser

Smart client-side apps bolster services-oriented computing

THE INTERNET'S SERVERCENTRIC model is demonstrating itself too slow and too static for the medium to advance beyond a document warehouse. Static forms are persistently cornering server resources. Round-trip server calls hog bandwidth and hinder the user experience.

Emerging distributed services-oriented computing models further call for faster, smarter workflow than is attainable via browser-based DHTML and JavaScript.

Rich Internet applications aim to smarten up browser-based interaction by making better use of client-side processing capacity. By leveraging tools such as Java applets and client-side databases to greater processing advantage, rich Internet applications improve functionality while reducing the number of calls to the server for information.

Screen updates, for example, need only include data changes, not an HTML document's entire page presentation. Localized data and processing allows users to perform tasks such as data mining and modeling with a higher level of interactivity.

The process helps to reduce network traffic and server loads, and allows administrators to maintain centralized management over their applications, keeping deployment and maintenance costs in check. And the decidedly peered architecture of rich Internet applications is providing a path for efficient application delivery within the highly distributed Web services world.

In the rich Internet application space, there is certainly no shortage of vendors vying for attention, although the degree of capability and sophistication of each varies.

The one commonality among all vendors, however, is the mission to create a smart-presentation layer and rely on localized processing to streamline interaction. In general, the approach is not a replacement for existing Web or application servers but rather an architectural supplement.

Whereas some companies such as Esual Software approach the effort with a focus on improving presentation, others are aiming to improve functionality in application delivery with entire workflow frameworks.

Notably, Fourbit Group, a Florida-based company with a forte for wireless infrastructure, is walking on sunshine with a recently inked a deal to bring its task-driven Fablets, a smart client platform, to Sun.

Also emerging from the vapor is Digital Harbor, hoping to make a splash with its Professional Interactive Information Environment, a unified interface for enterprise application access.

Macromedia is also a long way from its early days as a designer's presentation tool. Although technical limitations remain, the latest MX line boasts modest provisions for Web services and, with market penetration in the high 90th percentile, portends an opportune delivery mechanism for enriched applications to customers.

The two current standouts among the business marketplace competition, however, are Altio and Curl, each with extremely well-developing platform portfolios in hand and actual run-time environments available for examination.

Although their methodologies vary -- Altio uses an applet-servlet model and Curl a JIT compiler approach -- both produce solid means of improving client-side interaction to a wide array of device types.

Distinguishing these companies are other factors, such as Curl's required client-side language and hefty fat-client download, making it difficult for developers to hit the ground running and potentially limiting its usability outside of b-to-b scenarios.

But with that learning curve and overhead come some solid features, including a decently provisioned IDE and security model implementation, as well as good support for XML, multiple media delivery types, and compression capabilities that extend Curl's range of usefulness.

Altio is currently in beta with a forthcoming release of its AltioLive platform, slated to improve Web services integration and device-delivery offerings.

The specific direction these richer, browser-based applications will take remains to be seen. This inaugural wave of vendors is clearly struggling to conceive platforms with enough value potential to entice enterprise-grade customers to fork over hard-earned budgetary dollars.

For the large number of companies that spent the late '90s invested in retrofitting Web front-ends onto their enterprise applications, the dubious value proposition may appear to overshadow the merits.

Admittedly, many of the capabilities from today's innovators may ultimately take root directly within next-generation Web browsers, or they may be embedded natively within operating systems and application servers.

The risk reward here, though, is favorable. And in due course the capabilities coming to the forefront of rich Internet application development portend great promise for improving the sophistication and interaction in data and security-laden transactions.

By maintaining a focus on important features, such as an interdependence on XML, adoption of Web services and workflow mechanisms, and a moderate learning curve for developers, rich Internet application vendors will ensure bonding with corporate IT ideals and enable developers to bring to market more useful applications in less time.