Struts 2 may have rendered Struts 1 "obsolete and inadequate", but Struts 2 is strong candidate to consider against Spring MVC, Seam, and other current options.

Standards, standards, standards. WebBeans, JSF2, JPA2, Servlet3. Well, when they will become available ;-) And if they aren't perfect then surely WebBeans2, JSF3, JPS3 and Servlet4 will be!1!

We use Spring (not spring web flow) in our java web applications, and we love it. I'm not sure what "standards" it deviates from, unless you mean JSF, and for that we don't care: JSF is a mess in our opinion, complicated, overbearing, and it doesn't solve any problems it claims to.

JBoss seam appears to be JBoss's attempt to copy Spring's model and popularity; personally I don't see why you need to go with them unless you like having your entire stack from one vendor.

What about Tapestry 5? It's not released yet but it does show a lot of promise.

Branden -

I would assume that the standards Spring deviates from is the overall JEE spec, not just JSF. Which isn't necessarily a bad thing - Spring afterall emerged as a replacement for the the nightmare that was J2EE - but it's perfectly true to call it non-standard. It may support many of the individual JEE components - JPA, etc - but has a completely different way of gluing everything together.

Incidentally, Spring and JSF are hardly competing - there's a little overlap with Spring's web layer, but the two complement each other pretty well. We use Spring for the logic and persistence layers, while using a fairly thin java layer for the UI-specific stuff. Works well enough - I don't find JSF quite the failure that you do, though I certainly agree it has it's problems.

Stripes is a good option, very fast and near to 0 configuration

I've seen "Struts is obsolete" posts for a couple of years now from those enchanted by the latest and greatest, and--with due respect--this is merely the latest and one of the least compelling.

None of the RIA technologies I know render Struts obsolete. Certainly, as the market swings (no pun intended) towards richer interfaces, Javascript inevitably becomes more important, but on the whole, Struts is an MVC framework, which is still highly relevant, and Struts 2 in particular offers a wealth of tools that facilitate webapp development and organization. It includes, for example, workflow configuration, advanced templating, type conversion, validation, a rich set of control and form taglibs (including, yes, Ajax-enabled tags), and a host of third-party plugin's for things like reporting, JSF, and Spring.

Spring is certainly very popular (J2EE noncompliance notwithstanding) and offers its own framework, but I believe Struts is a stronger and more coherent framework for webapps--Spring's strength is in a deeper, front-to-back integration sense.

And, since I have development (and debugging) of JavaScript, I'm all for tools--like GWT--that can standardize the outputted JavaScript and provide Java-centric runtimes to facilitate development. But this still isn't a substitute for a proper MVC framework.

PS: I agree, Sun should have buried Swing a while ago, and JavaFX seems half-baked, for now anyway.

IMHO Stripes, Wicket and Tapestry 5 are the best options for the web client side, for the middleware Spring and Hibernate.

Wicket and Tapestry 5 are Component based frameworks very powerful stuff in there and also near 0 configuration.

Stripes my favorite as Mario Arias said it is a MVC as Struts but near 0 configuration, fast and very very easy to use, more easy than Struts 2.

Colonel35 why buried Swing, I love Swing it have it place if you dont like Swing there is another options as SWT or QTJambi very good also GUI frameworks for Java.

JavaFX it is mixed feelings for me I will adopt it if integrate well with Java code as JavaFX script for declare the GUI and use Java code for the events if doesnt work like that I will not adopt it, Still very young to talk about it lets give them time.

Alright this article is about web frameworks not GUI client frameworks my choices still are Stripes, Wicket and Tapestry 5.

My experience with component-based frameworks (Tapestry and JSF) has not been good. They generally feel like they're trying to force a Swing paradigm on the web, which just doesn't work. Wicket may be different, but the best option (IMO) if you want to code your UI in Java seems to be GWT.

In the action-based framework world (my preference), Stripes seems to be the best framework out there, in terms of borrowing good ideas from Ruby on Rails. However, I work for a large company and large companies are leary of new, smaller, less-proven frameworks. Which is one of the reasons WebWorks never took off vs. Struts.

The irony is that WebWorks is now under the Struts 2 brand, which makes it much easier to sell a technologically advanced framework in a corporate environment.

Consequently my pick is Struts 2: it offers the best balance between solid, up-to-date technology and corporate viability.

IMO, if you want to leverage the Java platform/libraries, AND enter the world of agile development, maybe you should take a quick look at Grails.

Grails is built on rock solid foundation (Spring Hibernate Quartz)
Groovy is a breeze to learn (for a Java dev that is)
All of the Java back-end stuff can still be easily integrated

The Grails community is growing fast + there is a conference in London in 2 weeks. Check it out.

JS.

I am personnaly using JSF with IceFaces and Facelets.
JSF itself doesn't really do the job, but the 3 of them together satisfy all I need : AJAX components, templating solution "à la Tiles", integration of J2EE security inside IceFaces tags, Spring Integration...

Kyle Adams, I tryed GWT but my experience with it was bad, It generate a bunch of JS files and I got lost in all that mess. I dont really like code generation tools as GUI builders they tend to generate very messy code. I read GWT 1.4 it is better and more optimized JS code but I would not like all my Web UI javascript based, I like more Wicket aproach. About action frameworks when I saw Stripes I got very excited about it, also I checked Struts 2 it is not bad but I liked more the simplicity of Stripes and have all the features. I wish Stripes join the Apache foundation so companies can take a look at it and embrace it more in the corporate.

I love the Web...
I've got two SUN certification before come to the tought that Java is not a 'good' choice to develop WEB appllications.
Now you have so many MVC frameworks that it seem ridiculous to have to learn them all.
The problem it is not the language, but the standards released by SUN: EJB are uselessly complicate, SOAP WS implamentation a real pain, servlets and JSP do not come with any useful specs (such as validation, integrated front controller pattern), JSF do not follow any W3C standards (old and verbose XHTML code, obstrusive javascript), the annotations bad news (no configuration but code expressiveness lost forever).
The final result is that you see a great number of 'BAD' coded Java EE applications out there... and i think it is not developers fault.

I had a similar article a couple of months ago (link below) , and it's good that you come down a make a recommendation.

My key criteria is for choosing is not technical (as most of the options you mention are pretty good), but commercial (if I learn this, will there be work for me?!)

Paul

I completely agree with you, Struts is a technlogy of the past and Swing is great but underused.
JSF is too complex (ever tried to make you own component?). JBoss Seam makes it less verbose (no XML config files ala Struts) but it's not enough.
GWT is impressive, but is all this DHTML stuff the right way to do RIA? I expect something better to come out from Flex or JavaFX.
After Struts what? Stripes!!!

Right now the hot stuff to get jobs are :
1) Classic Struts, 2495 jobs on Dice
2) JSF, 1471 jobs on Dice and rising

all the rest of web frameworks are below 80 jobs on Dice. So it means the Next thing is JSF but remember JSF it is a specification we need frameworks on it as MyFaces so on.

Technically as I said I like more Stripes, Tapestry 5 or Wicket. Those are on a niche but I would prefer to work on those niche than follow the wave as always everybody do. I prefer technical than comercial. Because that way I can work on the things I love to work on.

Mike Costa, I see Java perfect for Web development what else you suggest? Ruby on Rails, Php or Asp.net?, Asp.net it is a disater, Php its messy and ugly language, Ruby on rails it is ok but I feel Java more Strong for enterprise web development check out Stripes it have many ideas borrowed from rails. If you dont like EJB specefication no problem in Java land there is options you can check Spring and Hibernate it really rocks and intagreate well with Stripes and you can have a full stack.

Did you do any real development using JBoss Seam and JSF?

Yes, i've already used both SPRING than HIBERNATE: powerful tools but you have to learn another 2 frameworks to do standard things (GAILS abstarct their use anyway)...
Another problem with Java it is that the language is too verbose to experss simple things...
Take the reflection API: if you have to call, say, a getter method over a private attribute, you finish down writing 10 lines of code or near. In Ruby the same things can be done with no efforts.
The same are with RegEx API (and i like so much regex).
I loved Java, but now it simply seems to not respond to my love with the same enthusiasm...
I've started to develop in RoR and must admit that the framework is impressive (but again far from perfection). The most notable difference is that you are much more productive and it is less difficult to do 'bad' coding.
Do not stick your fingers into your ears: the rising of GROOVY, GRAILS and JRUBY must be a warning for all Java developers.
My toughts for the future of Web Applications?
60/70% of logic will be into client scripting code (say a more powerful version of JavaScript), with a 30/40% of routines and DB access performed by a (much lighter than Java) server side language (Ruby, Groovy, who guess?).
Stay tuned...

I use Apache Wicket

...
Mike Costa, I see Java perfect for Web development what else you suggest? Ruby on Rails, Php or Asp.net?, Asp.net it is a disater, Php its messy and ugly language, Ruby on rails it is ok but I feel Java more Strong for enterprise web development check out Stripes it have many ideas borrowed from rails. If you dont like EJB specefication no problem in Java land there is options you can check Spring and Hibernate it really rocks and intagreate well with Stripes and you can have a full stack.
Jeff | October 5, 2007 04:06 AM

Why do you say Asp.net is a disaster?

maestroQC, Yes I have tried JBoss Seam it's a good framework but I'm more confortable with Spring/Hibernate/Stripes.

Mike Costa, I respect your opinion and maybe you are right but I'm not so fan of dynamic scripting languages. Maybe in the far future the next best thing will be a dynamic scripting language on top of the JVM but as I said I don't go where the masses go. I think Java will have it place for a long time and if I want to switch to another language I will look at Scala it is very impressive language with impressive features and it can reuse all the Java api's.

Bobby, When I started with Asp.net it was version 1.1 it was so limited and then with version 2.0 came out it changed many things and yes improved but in the same time I have to relearn many things again as always all Microsoft products change without thinking for their users and I dont trust anymore Microsoft and it is only one vendor blah blah you know the story of M$ plus the markup of Asp.net is very messy similar thing as JSF using JSP.

I think we should not get obsessed by these Frameworks.These 'Frameworks' come and go every now and then.Basic programming remains same.Right?
The giant software systems/prodcucts have been built oevr the years without these fashionable frameworks.Don't get swayed by these frameworks.

Go Ext and keep the jsf. http://www.extjs.com Well engineered, cross browser compatible and an awesome look as well. Don't believe me, go see for yourself. There are some using it with GWT and more.

I duno. I'm rapidly becoming a fan Rails compiled with JRuby. There is something about cranking out simple stuff really quickly and having the power of the full JDK at your disposal that just makes me giddy.

I'll say Jython and got a new release 2.2. Python is better than Ruby.

OH MAN, SEAM IS IT!!!!

Bow before ZOD, all other frameworks (you too rails). The things you can do with Seam you would have to spend ages configuring elsewhere. RichFaces in Seam makes AJAX painless. Integration testing is also heavenly with it's component framework. I love it. For those not using Seam....your ignorance is your downfall, and I mean that in a nice way.

JSF is the worst solution for the Web. It is all POST. That alone should keep you away from it. Most of the Web is GET.

Also, JSF's automatically generated JavaScript is a Web designer's nightmare. Our senior Web designer hates JSF because its client-side mess.

Most of the Java developers on our team prefer Spring or Struts 1. However, we are forced to use JSF because our architect chose it. She has no Web background. Go figure.

I'm not sure I understand this:

Spring is a great choice if you already use a Spring server side, everybody in your team knows the spring framework well. However, Spring deviates from the standard JavaEE in many ways and I would suggest sticking to the standards.

You are not forced to use any other part of Spring if you want to use Spring MVC and it is a very good web framework. It is easy to learn and quite honestly if you learn how to work with it you pretty much already know how to work with the rest of Spring.

In what way does Spring deviate from Java EE? You say there are many ways yet you don't provide any examples. Spring is widely used and accepted so I'm not sure what you lose by not "sticking with the standards". In fact since Spring provides such nice interfaces to parts of the standards and makes them easier to use I'm not sure why you'd not want to use Spring.

How about a Java port of the following?

http://code.google.com/p/barebonesmvc-php/

It's just 60 lines and I actually came up with the general concept (applyInputToModel(), applyModelToView()) working on a Struts app for Chuck Cavaness. I've also implemented it for mod_python/python_server_pages, but rather than additional ports of my once-and-for-all solution to MVC (for the web), I'm moving on to solving ORM once-and-for-all; I just have to implement the solution that hit me last Saturday morning when I woke up.

Carry on

I agree JBoss Seam is impressive but it is a full stack and we are comparing web presentation layers, with Seam you can switch web presentation layer as Wicket I think they are working on it instead of JSF. For me is Spring/Hibernate > EJB3, Spring doesn't need an application server it is very lightweight, I can apply Spring even for Desktop applications. Spring is very easy to use and I can choose xml configuration or anottations, With EJB3 without anottations it's a nightmare. Actually Seam without anottations you can get lost in a sea of xml, still complicate without anottations and the glue code that provides Seam.

The point of Spring framework is Middleware, SpringMVC it is just as an addon to the framework so it gives you out of the box a web presentation layer but it is not compulsory to use it you can use any Web presentation layer with Spring as JSF, Wicket, Stripes, Struts so on or heck you can use Spring for the middleware and front-end Swing in a desktop application. Spring Webflow it is also good addon and you can apply it to JSF or Struts so on.

Jeff,

FYI, Seam doesn't need an app server either.

Ditto Dan - Seam is the way.

I'm on my third Seam-based site (all based on .war's rather than .ear's) to date. It's simple, straight-forward, and has a ton of functionality built into its core - integrated security, PDF's, templated email, seam-gen, etc. And just as important is the Seam community, which is innovative, productive, large, growing, and friendly.

PHP with PEAR component libraries.

Guy Mac, PHP sucks Big time my friend, It is the worst designed scripting programing language in History of IT. I prefer to use Perl any day than PHP. Better go with well designed Languages as Python or Ruby or Java.

Just spent the weekend at NFJS in Atlanta, and I've seen the future: Groovy and Grails.

I got press-ganged into doing web development about a year-and-a-half ago. I've been using Spring MVC, with Hibernate on the back end. For the stuff I do - mostly CRUD websites - Grails would provide a massive boost.

For the first time in a long time, I'm actually looking forward to a web ui project.

I prefer Swing myself.

It's far more cost effective, imo, to bang out a Swing solution as opposed to a web based application. There's simply too many what-if's and technical gotcha's to support multiple browsers. And browsers are slow. And they don't interface with the host computer particularly well.

I used JSF/Facelets for a couple of years and wrote a few complex applications with it. It's far more complex than it needs to be with it's 7 step lifecycle and all.

Hi Bob,

I'm with you on the Swing thing. The first Java apps I wrote had AWT front-ends (I just missed the 1.0.1/2 event model. Yay!), and writing Swing feels like getting back on a bike: you never forget how.

That said, the web is where the action is at the moment, and my wife likes to receive new shoes on a regular basis. ;^)

If you like Swing programming model, you will love Wicket it is very similar programming model for web application development. You should really check it out, It is very different to JSF it is more close to Swing style but it is a very nice framework and easy to use.

Patrick Right now my favorite web frameworks are Stripes and Wicket. If I have to use JSF or Struts it only because jobs. You can check it out Wicket as I said if you like Swing style programming model or Stripes it is the more simplest, elegant and easy web framework I have seen so far.

SF is the worst solution for the Web. It is all POST. That alone should keep you away from it. Most of the Web is GET. Also, JSF's automatically generated JavaScript is a Web designer's nightmare

That depends on the JSF implementation. Try the RI instead of MyFaces. It does not generate JavaScript. Also, not ALL is POST. You choose yourself.

We are talking about Web or what? Swing? Applets? Flex?
To me the Web have to stay with the W3C standards, not the SUN, IBM, Adobe or Microsoft ones.

The responses are so varied, that it is clear most people in the Java community are unsure about what next after Struts!

My response to some of the comments are as follows:

Dan - Yes Struts 2 is a good framework but it is no way as popular as Struts. Its still an action based framework and its new contribution is mostly around support for AJAX. For those who liked WebWork2 its a blessing but for those who prefer the event driven component frameworks (like Apache Shale) its still away from the mark.

Many teams that moved on from Struts did not adopt Struts 2. A blog post by Don Brown had stated what Struts 2 was brining to the table and if you haven't already read it you can read it now -
My history of Struts2 -
http://www.oreillynet.com/onjava/blog/2006/10/my_history_of_struts_2.html

Nik - You are right its a pain to keep up with the standards. especially when the Java community loves to create so many of them . However, its always beneficial to go with the standard than not.Invariably container vendors do provide support for the standards.

Branden, Jason, Simon, Patrick and everybody else who spoke about Spring, here is what I have to say:

Spring is a great framework - there is no two ways about that! Spring emerged at the time when J2EE's EJB model was cumbersome and painful to deal with. It provided a lightweight implementation for server side beans. It made programming easy for those who were using EJB then. It also adopted a lot of good OO principles in general for building applications. At that time some of the OO principles were compromised within the world of J2EE because of the constraints of the specification and its implementation. Spring was and is an alternative to J2EE.

On the enterprise Java front the story has changed. J2EE has emerged from being a clunky and heavyweight framework to a lightweight and easy to use framework. Especially with EJB3, the server side beans are much much simpler than its ancestors. They are also far more elegant. Now the big question is - should one use the standard JEE (its no more J2EE, as per Sun!) or should one go with Spring? If there are frameworks that are built on top of standard JEE and provide all (or should I say most) of the features that you need to build web applications, would you not prefer that?

About Seam copying Spring success story -
Because Seam is more recent as compared to Spring this statement could be accurate. However, there are two big differences between Seam and Spring -
1. Seam reuses and integrates, JEE and other JBoss products to build the entire stack, spring pretty much re-implements the whole stack.
2. Seam is built around a stateful model whereas Spring is built around a stateless model. This does not mean that Seam does not support stateless interactions or Spring cannot implement statefulness.

Lastly, as far as web frameworks go, it was rightly pointed out that we should speak of Spring MVC and maybe Spring WebFlow and not Spring as a whole. You can use Spring MVC without Spring but then you are left to do some of the plumbing. The alternative is to adopt the whole stack. If you are already heavily invested in Spring, this is an obvious choice. However, if you don't have that much Spring in place do you want to learn a new model, when you may already be aware of many things JEE. Personally, I like Spring but also hate its XML configuration files, ironically it reminds me of the old EJBs.

Manuel,
Why do you say POST is bad and GET is good? The two HTTP methods have a different purpose. For sure a lot of the web resources are accessed using the HTTP GET method. However, that does not make POST a bad method. JSF does rely on POST. Seam defines the concept of a conversation, which is smaller in scope than a session. There can be many logical units of interaction within a session. Each of these unit of interaction is a conversation. Seam treats each GET call as the beginning of a separate conversation, unless deliberately implemented otherwise. This also means that as far as the POST methods make a call within the same JSF lifecycle they are in the same conversation. This is useful and logical. What say?

Stripes, Wicket and Tapestry are good frameworks. In fact facelets which is slowly and steadily becoming the defacto JSF view description technology gets its inspiration from Tapestry.

However, they are not extremely popular. Interestingly, those who use it swear by it. Maybe because there is so much to choose from or maybe because there is no strong force like JBoss, the Apache foundation or Interface21 providing the necessary thrust. A few months back, Ed Burns,the JSF spec lead was appealing to some of these guys to join hands in creating the JSF2.0 specification. Maybe they should actively participate.

JS, Vance, Patrick, Mike, John and everybody else who spoke about non-Java frameworks - Grails, Rails....

Java is as much a platform as a programming language and certainly frameworks, especially web frameworks, based on programming languages other than Java are becoming popular. Scripting and dynamic languages are always easier to learn and easier to use and they may become the dominant language one day. For now we have 10 million Java developers and rising and maybe the preference will remain with something Java.

Hi Shashank,

Just FYI, Groovy/Grails are built on Java, and a lot of other stuff (Hibernate, Spring, Quartz, etc.).

Patrick,
You are right Groovy/Grails are built on Java. As I was saying Java is now more of a platform rather than a programming language. It was done as a JSR. I am on a bunch of expert groups and have some visibility into how it evolved. The syntax of groovy can be made pretty much the same as that of Java but purists may then argue that it looses its Groovy characteristics if you do that. Groovy has some newer features over Java, for example it has native support for properties like C#. Groovy objects can be used as Seam components too. For that matter JRuby and Jython let you run Ruby and Python is Java. Yes they are not as native as Groovy, but you can very easily use TurboGears or any other Python framework and call it from Java objects and integrate it with Java libraries using Jython. Quite a few years back I wrote a Python GUI library that, where you could write the code in Python and run it as a Swing application. Also you could call the Python libraries from within the Swing application.

Anyway Groovy is good and more Java (I mean the programming language now) like than anything else but can it be the mainstream choice of Java developers.

I'm not a bigot but I'm very confortable as how it is Java now. I don't want Java become the C# footsteps, C# is to bloated and complicate right now IMHO. Java still simple and easy of use, yes it is sometime little verbose but I like it since first appeared in 1995. For the folks that likes more advance or complicate (depend POV) programming constructs as closures and such I will suggest use a language as they are suggesting that have that stuff out of the box as Groovy or Jython on top of the JVM, Me I stay with Java, I don't like dynamic languages.

Sashank Tiwari, I think we all know what is the next Struts because the popularity it is growing more and more and that is JSF .But also I like choices as for web presentation Wicket or Stripes as middleware Spring, yes EJB3 is the new standard and it is easy to use but I like choices and as many people say use the right tool for the job. Maybe sometimes is better to use EJB3 and another times Spring, Im open to choices and tools but for me Spring has been a great software and continues to improve.

Don't commit to only one solution as only one web framework. for example me I use Wicket and Stripes, Wicket is great for complicate desktop like web applications becuase you can construct very detailed UI components with it and in the other side I have Stripes to build High trafic websites but easy and fast. Plus I will say it I can do also JSF but that one right now it is just to get a Job heheh.

I love Java the language, the JVM and all the choices and frameworks we have on it. I like freedom that is why I dont do anymore Microsoft products development. I like Java freedom and now more it is GPL.

Nice post !! In my opinion dont have any better framework that those wich are create with specification like JSF !!! I dont have dought that JSF will became what struts 1.x was in the past. I've been used JSF since 2006 !!!

I recently did my own little evaluation on whether there is a better alternative to JSF and you can find the results here:

http://ptrthomas.wordpress.com/2007/05/14/a-wicket-user-tries-jsf/

I'm still not convinced about Seam especially given the conclusions I reached regarding JSF. Guess I have to get around to doing another comparison soon...

All the best for the book you are writing on Seam ;)

Don't underestimate Tapestry. Tapestry 5 is a pleasure to work with even in its less than finalized state and it is extremely easy to create your own components or extend existing ones. No TLDs to deal with, and only minimal XML configuration to set up (to which you can then generally forget about). Seat of the pants, it seems blazingly fast by comparison to JSP based frameworks.

I would add my voice to the many developers who have mentioned Wicket. It is a very "Java" solution - there's nothing to it but Java and HTML markup. The clean separation betweeen markup and code is very appealing.

Rolland, I'm agree with you, Tapestry 5 it is really a great framework, Do you want to do JSF 2.0?, why wait take a look at Tapestry 5, great features, very modern framework. Lately I'm doing some demos with it, I like it a lot. Just the Docs have to improve little more and the tutorial is not done yet, The reference it is ok and need a Tapestry 5 in Action book.

"Apache Struts is the most popular Java web application framework till date. However, newer developments have rendered it both obsolete and inadequate"

WHAT?!?!?! what planet you live on man???? That is such BS.

Peter,
On planet earth among Java web developers there is absolutely no doubt that Apache Struts has been the most popular framework so far. Maybe you want to check with Apache on the number of downloads. Maybe you want to get data from all types of job sites to understand what people have been using.In fact its highly unlikely that anything can become as popular as Struts, partially because Apache Struts came at a time when there weren't many competing frameworks and it solved many of the problems JSP/Java Servlets developers were struggling with at that time.

Now about the part that its becoming obsolete, things are fairly obvious. Certainly its still in use at many places and many may still adopt it but With AJAX/RIA gaining momentum developers are moving to event based models and moving away from action based MVC.

Hope this explanation helps!

Peter...

He's on planet Earth. What planet are you on? Both of his statements are very uncontroversial.

If you like nice OO design and implementation then
Wicket + Spring + Hibernate
is the best choice!

Seam is real, it works and it will be maybe THE framework for next few years.

This is an excellent question: after struts-what?

We wrote our application using struts and find the choices mindboggling. The java world presents a mind-numbing array of products, frameworks, tools, buzzwords,etc. Many new fashions fall by the wayside.

As developer maintaining and expanding an existing application, we face two questions:
1) For new functionality what "new tools/frameworks/technologies" should we use?
2) Do any new tools/technologies present so great a benefit that we should migrate all our existing jsp's, struts actions, etc.

#2 is a lot of work.

There are so many new technologies its' hard to keep abreast.

OnJava should provide a 'bellweather' (e.g. like a top 10 list) of most popular technologies (whether by downloads or interviews with CTO's of leading companies or something to help sort out the acronyms and buzzwords)

As long as the application being developed does what its supposed to do, any of these technologies are good. They are a means to an end.

This is really an interesting issue.

Running the Googletest gives interesting results.

I've searched results within the last month on google, counted the results and displayed the results here:

http://mortenmatras.blogspot.com/

My answer is relatively straightforward. JBoss Seam.

Most other frameworks are like a nice polished dining room table with a some foot wide holes in it. They don't offer the whole package, but what they do offer looks good. Seam offers the full polished table without you having to fill the holes.

JSF too complex? Not at all. Use Seam which incorporates facelets and see how many flaws you find in JSF. Still find it too complex? Eh, go find another career, JSF isn't that hard, and you don't need an IDE to build the page.

Don't like the idea of a component model, fair enough, some people prefer assembly or procedural programming over OO programming. If HTML is your bag, go nuts.

Remember, RoR rose to power for a reason, because the java world wasn't fullfilling a need. Seam fits that need. Create a blog in 15 minutes with RoR? Heck, I can code up a full Ajaxified CRUD app, complete with a DB from scratch, multi-window and back button conversation management that will scale effortlessly over a cluster in 15 minutes, not to mentional all the points are there to plug in authentication, security, and javascript remoting to my session beans.

Spring is an old hat proprietary single vendor platform which wowed the world (and rightly so) with its implementation of dependency injection. However, Seam will be the new standards. Someone already noted the number of JSF jobs is growing like crazy. JSR 299 (web beans) will standardize Seam making it pretty much a standard application stack from end to end, not to mention having its own IDE based on eclipse. Once the whole Seam stack is standards based, where do you think the job numbers will go from there?

Good analysis. I have same debate in my team that what should we use for our web project(Its an ERP project for manufacturing industries).We have three developers experinced in struts, but as a number of frameworks have entered in J2EE market so it created a debate here.
we are also looking for some other powerful and good frameworks like spring or jsf.Now after SEAM 2.0 release,we are also considering it.as a architect, i ahve to give recomendations.
So i am in bit confusion so what you suggest? (i am afraid that you will find it difficult to suggest as you haven't full picture of my requirments)

What do you think of General Interface from Tibco? It has a BSD license or a enterprise license. It is reviewed very favorably at http://akamai.infoworld.com/Tibco_General_Interface/product_70328.html?view=0&curNodeId=7

They are developing Direct Web Remoting to push data to the client. I saw tutorials here:
http://www.theserverside.com/tt/articles/article.tss?l=UsingDWRwithTIBCO

It looks good, how do you think this compare with Seam?