JBullet 1.0, a convention-based web framework, has been released. It mirrors the Ruby on Rails url capabilities very closely, having a default method for the controller. However, it uses Java annotations over the Rails types for controllers; the documentation is simple, although still fairly rough. From the sample "Hello, world" application, you can see how the controller implements urls://Listing 1.1 //File HelloController.java package controller; import org.jbullet.controller.BaseController; public class HelloController extends BaseController{ public void index(){ renderText("Hello World"); } public void wow(){ renderText("Hello from WOW"); } }//End Listing 1.1Calling http://host/controller/Hello/wow would invoke the wow() method, and http://host/controller/Hello would use the index() method by default. Dependencies are injected through annotations, and models are as well (through @Dependency and @Bean, respectively.) Presentation is managed through a custom tag library, that provides close analogues to the normal HTML tags.

This seems to me the same thing as Stripes (http://stripes.mc4j.org/confluence/display/stripes) does. Where stripes seems to be be more mature and well documented and finally did the job for me.

Okay, I'm impressed by Stripes. I have looked at many, many web kits through the years (Struts, JFC, Spring, Grails, etc etc) and Stripes looks like it actually IS worthwhile to use.

Yet another framework in the already clouded frameworks.

Seriously. I have a crazy idea for those of you itching to create a new framework. Go find a framework that, you know, already exists, and figure out how to do your stuff with them! Because, while a million Java monkeys working on a million webapp frameworks might not get very far, a million Java monkeys all trying to make _one_ framework better might get very far indeed... Personally I like WebWork, but hey, pick any of the handful of existing frameworks that people are actually using in production applications. ;)

The current popular frameworks are overly complex, clumsy and barely usable.

The current popular frameworks are overly complex, clumsy and barely usable.

All 458 of them? ;)

The current popular frameworks are overly complex, clumsy and barely usable.

All 458 of them? ;)

Yes!

The current popular frameworks are overly complex, clumsy and barely usable.

All 458 of them? ;)

Yes!

I agree. For me at least all frameworks that still use JSP/Taglibs don't cut it: We work with webdesigners a lot and I want to allow them to do their work freely without any illegal HTML and/or logic in their pages. That leaves only a couple of frameworks. And for me Tapestry comes out on top. But still... I think things could be so much better. I think people should keep working on their own frameworks. Someday someone will come up with something that blows all other frameworks away. I started working on a web framework as well, but that also probably won't be the one to rule them all :oD Besides... you learn a great deal by writing such a framework and if you share it, maybe someone else will find it usefull. You don't have to read up on each and every framework if you don't want to. Just keep using what you like and when something is very good, you're gonna hear about it soon enough anyway.

The current popular frameworks are overly complex, clumsy and barely usable.

All 458 of them? ;)

Yes!

So who do you blame for that? What's your contribution to fixing this? Are you involved in any initiative? Did you ever try to even write down some specs of how you think a good framework should work? And do you expect a framework that is not overly complex (do you mean in use or in iternals) to solve difficult problems for you, or do you want something that works quickly for easy stuff, but that will demand serious hackery if you want to do more complex stuff? I'm really curious here. What do you think is a good framework outside of Java? Is there any? Any you actually used?

Like most people who reject the major frameworks I only got investigated the first 12, or so, I came across. These seemed to be derivative and seemed to offer fairly similar development strategies. Some of the ones I did try blew up after a day, or so, of development due to bugs and limitations. So the main reason I think that there are so many frameworks out there is due to the large amount of code that is available to create a simple persistence and web interaction layers. The native JavaSE libraries, Jakarta commons, JDBC, JOTM and the like are great tools. So if you reject the thrust of the major frameworks (XML configuration files and Javabeans in my case) the siren call to create your own framework is almost irrestible. I blame noone for the huge spread of framworks. Everyone is trying to write good code and share their experience. So I suppose we should just smile and take a look at the latest killer framework. :) I'm still thinking about what I liked about the frameworks I've been using in the last 6 years of Java, Ruby and C++ development. It always seems to be a mixture of liked this, hated that; this was solid and that was brittle. Roll on the evolution!

There are two models of progress that we tend to see in IT. One is the model of incremental technologies, this is similar to the breeding of dogs/horses where the aim is to continuously improve on previous technologies. You start with a good technology and try to make it better. This is most prevalent in the commercial (read software house) sector. It's strength is that (roughly speaking) every generation of technology is already of a good 'breed'. It's weakness is that you are ultimately in-breeding and innovations tend not to appear. The second is the model of evolutionary technologies. This is similar to the process of evolution. There is no aim but the idea of 'better' and 'survival'. Each product is trying to do a job better (either a new job or an old one) and it is trying to survive against intense competition. This is most prevalent in the open-source arena. It's strength is the ability to make quantum leaps and to find left-field solutions to existing problems. It's weakness is the huge multiplicity of technologies formed and the _average_ quality level being much lower (although the peak may be much higher). Since we live and breath in both worlds as Java developers should we be suprised or disappaointed by the several hundred frameworks that the evolutionary process requires to produce the killer frameworks which outstrip the incremental process (e.g. Spring outstripping J2EE etc.). The idea is a good one, best of luck guys.

In which way is //Listing 1.1 //File HelloController.java package controller; import org.jbullet.controller.BaseController; public class HelloController extends BaseController{ public void index(){ renderText("Hello World"); } public void wow(){ renderText("Hello from WOW"); } }//End Listing 1.1 simpler or more useful than: <% // Hello.jsp %> Hello World <% // wow.jsp %> Hello from WOW ?

In which way is

//Listing 1.1
//File HelloController.java
package controller;
import org.jbullet.controller.BaseController;
public class HelloController extends BaseController{
public void index(){
renderText("Hello World");
}
public void wow(){
renderText("Hello from WOW");
}
}//End Listing 1.1

simpler or more useful than:

// Hello.jsp
%>
Hello World

// wow.jsp
%>
Hello from WOW

?

:-) I agree that the example is wrong. The other one where hello.jsp/wow.jsp is written is more easier!