I sense a change in the wind. Trends in IT are just as cyclical as they are in financial markets, and just as difficult to predict. I hereby stick my neck out and suggest that the powerful trend towards thin client "browser-based" applications is running out of steam. The browser bubble is about to burst.
As evidence I offer the following nuggets:
Browser feature slowdown
Browsers seem to have reached a threshold in terms of features. Sure there is lots of work to do to get browsers more compliant with standards of various kinds but I cannot think of any planned browser feature from any source that makes me say "Wow! I want that!"
HTTP-aware thick client software
HTTP is finding its way into applications quicker than applications are finding their way into browsers. I have numerous "thick client" applications on my desktop that know about the Web but are not browser-based. They can read resources over HTTP and they can write resources with FTP, Webdav, and so on. In a sense, these applications are "browsers" but their front-end's are not generic and critically not HTML based.
Earlier in the growth of the browser bubble, application developers felt the need to embed their applications directly into browsers. The height of this trend for me was a "browser-based application" of my acquaintance that required 80M-bytes of ActiveX controls in order to function in a browser. Not exactly a thin client application!
These days the trend seems to be to leave applications self-contained and to add web-awareness rather than embed them into browser containers.
This is a truly remarkably office suite that is web-aware and stores all of its data in a readily processible, XML form. One way to look at it is that it is a set of web-aware, XML-aware widgets from which highly customized browser environments can be constructed. The possibilities are intriguing.
Instant messaging, weblogging and P2P
Three very interesting application classes that are extending our notion of what "on the web" means. Although all three have traditional browsing technology somewhere in their feature sets, use of browsers is peripheral to the core of these application classes.
A critical aspect of web services is that resources are separated from HTML and thus separated from reliance on traditional desktop browsers. As such services proliferate, so too will thick client applications that engage directly with the web services, by-passing HTML based user interfaces.
None of these points on its own would carry the argument, but taken together, I think they may point to a change of direction in web application development.
Simply put, the boundary between an operating system, an application suite (such as an office suite for example) and a browser are about to get very blurred indeed.