Presentation

The most commonly used architectures for building application user interfaces (UIs) are thin clients and thick clients (or, as Microsoft prefers, “smart clients”). They are joined by an increasingly common third category, Rich Internet Applications (RIAs), which offer some of the interactivity and graphical sophistication of thick clients, with the browser-based deployment model of thin clients. (Although Microsoft, like Adobe and others in the industry, uses the acronym RIA, Microsoft prefers to call the category “Rich Interactive Applications.”)

Microsoft’s developer platform supports all architectures. In fact, it often has overlapping, competing products for a particular architecture, and the company steadfastly avoids giving long-term recommendations for which technologies are likely to continue to be developed and improved and which will be downplayed over time. (For a discussion of how to evaluate overlapping developer products, see the sidebar “Software Darwinism“.)

Thin client. Thin-client architectures allow a single generic client program to access multiple applications on servers. Although HTML displayed in a Web browser is the thin-client model most widely supported by application servers (in fact, thin client and browser-based have become almost synonymous), thin client also describes many terminal-based architectures, such as IBM 3270 terminals and Microsoft’s Terminal Server product. In each case, the bulk of the application processing takes place on the server and the results are sent to a generic client program using a standard presentation protocol. However, the introduction of AJAX (Asynchronous JavaScript and XML) programming is allowing thin-client applications to have increasingly sophisticated UIs. Application servers facilitate thin clients by integrating with Web servers and providing tools and APIs to connect HTML to business logic.