I think the next 6-12 month we will see an incredible buzz about web application frameworks – some on the server side and some on the side of the client:
OpenLazlo is competing with Macromedia Flex. for the so called “Rich” Internet Application market. I am somewhat sceptical about thie RIA-approaches. If you can establish a channel to deliver anything useful – fine. But these systems – while cutting development time – are extremely monolithic and they sort of hijack the user experience for you. And caring for the user experience is a differentiator in the market. I admit that Flex/Lazlo would provide a better experience often compared to using no interface toolkit at all – but generally I feel these systems are bloated and heavy trying to solve so many things at once. But maybe I am just not getting it right. Many internet applications would be “rich” if only they would be designed better and provide some sort of solution to a problem.
For the development of the backends on the server we see frameworks like Ruby on Rails, TurboGears, Twisted, Zope and so on. These can also really cut development cost by providing abstraction layers to common problems. These frameworks do a hell lot of things for a developer, but they usually have a steep learning curve and they also may have issues with reliability, performance and scalability. But from what I see, there are little options to avoid frameworks unless you are ready to invest the time in working around so many different gotchas yourself.
People were critizising AJAX to break the URL schema. I don’t think that is the case as long as you keep URLs as pointers to resources functional and constrain AJAX to improve the usability of a web application. I don’t want to have to reload a full HTML page or submit a complex form just because I unchecked an option.