Trying to follow…

Working with different types of widget library’s one gets in contact with Ajax libs. Personnally I think the difference is even not there. As soon as you get asynchronous calls to the backend (whatever backend that may be), we are talking Ajax. And a widget is a piece of code that works on its own and thus has his own method for calling the backend. Hence widget = ajax in a simple form. But ajax then… I thought it was a complex matter to get the communication with the backend going but these days I get in contact with a lot of frameworks that seem to get the job done. There is Google that has several intakes in the issue, like their gadgets, the mashup editor, GWT, etc. There is off course Mootools and Scriptaculous but they are more of collection of snippets. There is also Dojo, DWR, echo and the commercialy backbase, tibco, jackbe,….

But you have to make a distinction between server and client frameworks. The server frameworks have a server side footprint and a small client base (eg gmail), the client frameworks can  potentialy work offline but have a bigger client lib –> more downloadtime. But that is relative because today a lot of pages go beyond the 500KB and that is a lot of javascript.

Going offline is also a key component for the future, like google gears. That means that the server-side story is slowly going down and can be replaced by a client-service oriented architecture. Those services provide a readable URL interface to underlying systems with a atomic structure… True webservices without the problems, just loosly coppled systems. It also fits in the development of RIA, Rich Internet Applications, were the OS will be independent (or as some say the internet will be the OS).

It is a complex picture but in constant motion and as the story continues I wil be trying to follow it. So check back here for more updates and technical remarks. PLease feel free to comment on the issue…