As for the AJAX news, just the other day I was looking at the javadocs to the alpha version of the Jetty 6 API and saw the continuations class. It does look like a very interesting way to save on resource utilization at the server while providing a near real time feel to actions occuring on the browser. However, the article mentions only ONE use of AJAX techniques. Namely, the browser making requests to the server that don’t require immediate response, or that require response only if the status of some server attribute changes. The more popular (and easier in the sense that server modifications are not required) to use AJAX is to significantly reduce the data size of requests. This is how google and other sites are using it, to make periodic requests to the server for small data packets which are then incorporated into a standing page by some javascript. The overhead of the periodic requests is lestened by the significantly reduced data size which speeds up the overall browser experience significantly. Continuations would seem to be more useful where remote object calls are made, having a browser request trigger notification of the client when a server action is performed and possibly sending object data (not just snippets of xml destined for formatting in html at the browser) in the process. Still, continuations are not required to facilitate AJAX in Jetty or any other web server platform (any) if the data requests are small, the data delivered is small and the update periodicity is not too frequent.

Thanks Kevin.

AJAX applications will often send a request to the server with the expectation that that request will not receive a response until the server wishes to notify the browser of some event

I find that a little hard to believe since developers have to know they are hammering the server when they do that. There is a general understanding that the client app will need to initiate each request and that the server is not to be tied up.

Of course this also introduces problems since the server keeps getting tiny requests to check if there is an update. Theoretically the “continuation” idea could free up a lot of resources.

If this ever takes off I would expect to start seeing it implemented in a broad range of servers and not just for Java.