This is interesting to see the evolution of open source java based frameworks and combined that with trends' estimates raised by analysts (Gartner, Forrester;…).
 
So one of the latest hot topic is about the development and management of new Content Enabled Vertical Applications - CEVA ( http://www.gartner.com/DisplayDocument?doc_cd=137675).
 
But what does it really mean in term of implementation and application framework that such new applications could leverage onto (at least in the Open Source JEE world).
 
There always were lots of open source java based web development frameworks. Similar to the Fashion industry, developers also have their flavored frameworks and such tools are usually varying over the years.
 
Remember! Coming back a few years ago there were some tough battles with all the fans and critics of using one or the other Web Application Frameworks such as Apache Turbine (Is someone still using it?) or Apache Struts (and may others). Struts finally looked like having got the favor of the largest number of interested developers at that time. This was the first wave of web application framework.
 
A few years later the Spring framework hype took precedence on the Apache Struts trend. In parallel there was the heavy promotion of the new Java Portlet standard (JSR168) mainly done by the main IT industry players (IBM &Co) which all launch their own XYZ Portal Server. This was about the second wave. But it now looks like declining from its top in 2005 (cf: http://www.google.com/trends?q=spring+framework%2C+portlet)
 
Finally we are now seeing the apparition of a third wave of web application framework: Content Enabled Web Application Framework (CEWAF?). This wave is also leveraging new web 2.0 oriented "mashups" which are also taking precedence on portlets starting from 2007 ( http://www.google.com/trends?q=portlet%2C+mashup&ctab=0&geo=all&date=all&sort=0 ).
 
So what would be now the best way to rapidly develop 3^rd generation of portable, reusable, easily deployable, interoperable content enabled applications and focused on fostering social relationships?
 
One would then look now at the assembly of an ideal web content framework which would rapidly allow him to:
-         Easily manage its own local content objects without having to reinvent the wheel in term of content persistence, basic content management (graph and tree oriented kind of basic operations) or in term of content features which are today becoming more and ore commodities (properties, relationships, mixins,…). In the Java open source world, it looks like Apache Jackrabbit is a nice library to start with.
-         Do not have to take care of redeveloping all the RESTful, OSGi, Scripting facilities which are today quite commonly required and to just leverage a simple and ready-to-use framework. Apache Sling could certainly rapidly become this framework.
-         Easily consume content syndicated from other content oriented applications or from 3^rd party content repositories. The CMIS standard looks like currently the way to go and then in term of development the Apache Chemistry project.
-         Be able to leverage existing customer architecture and all their already deployed Corporate Portal Servers and be able to leverage the key strengths of such platforms (SSO; Personalization; cross-portlet communication…). In order to ease portlet development you can then start by looking at Portal Starter Pack such as the SUN one : http://portalpack.netbeans.org/
-         Be able to leverage the new social capabilities which are more and more often required today by customers by implementing standards such as OpenSocial (cf: Apache Shindig).
-         And finally work with existing Web 2.0 AJAX frameworks such as GWT while already having all the bridges in place with the aforementioned back-end libraries in order to get some immediate hands-on.
 
So will the future 3^rd wave web development framework an assembly of JSR168 Jackrabbit+Sling+Chemistry+Shindig Portlet?
 
The time will say it as most of these projects are quite new and have still not reached their maturity point. But how more simple it would be to develop a new content enabled application if there was already an easy to use and in-depth integration between all these initiatives. Developers would really be able then to focus on developing value added content oriented features rather than spending time trying to assemble a state-of-the art framework.
 
And finally, if this is about the 3^rd wave of java based web development framework, what about the 4^th one? Could we already envision what is will be?
 
AFAIC I personally think that the 4^th wave will add support for more Semantic Web oriented kind of features. Such web development frameworks will have to ease the work of standard content developers in order to more easily managed LinkedData.
 
So perhaps such a new generation of development framework will come out from projects such as the EU funded IKS initiative. Who knows! (and see you all in Salzburg end of this month in order to discuss about it)