1 2 3 4 5 Previous Next 73 Replies Latest reply on Nov 26, 2011 10:21 AM by kaczortrex Go to original post
      • 45. Re: Seam.Next discussion

        Miles Chang wrote on Oct 14, 2011 11:46:

        What about the Seam JMS/Drools module. Would they be discontinued?

        It's being moved into the JBPM project. There are a few issues we need to help solve first though, however, we've been swamped trying to get everything ready for the 3.1.0 release.

        JMS is still around, I assume you meant JBPM.

        • 46. Re: Seam.Next discussion

          Alexandr Areshchanka wrote on Oct 05, 2011 15:19:

          Alexander Kunkel wrote on Oct 05, 2011 14:41:


          I love Seams roundtrip developing CRUD applications
          - Change data model
          - Reengineer it into Seam Entities and views
          - Modify something and add additional functionality
          - Ready to deploy

          I only need Seam 2 plus
          - Flexible seam gen to invent some templates and own annotations to add further metadata to entities like OpenXava does
          - JSF2
          - JBoss AS7


          For spring and JSF 2 I found springfuse.com for crud generation.

          Do you really need AS7? What is so special in it, a can't understand still))) Just switch off tomcat logging and you will get same performance and startup time(actually what is done in AS7).

          If you'd like to use springfuse, go ahead, no one is stopping you.

          As for the AS7 / Glassfish / TomEE / any fully certified server, go ahead and roll your own using Tomcat 7. What you get when you use a fully certified EE server is that you know all the pieces work together. With plain Tomcat or Jetty you have to build all of that up on your own. I've seen quite a few posts on this forum alone about getting things to work with Tomcat such as JNDI, or JTA. There's so much more productive things a developer could be doing instead of trying to piece together a server, yet people continue to do it. We're a bunch of masochists. There actually isn't any official spec definition for a plain servlet container. The fact that it works still amazes me.

          Honestly though, if you must use Tomcat instead of another server, sell people on TomEE and use a full blown EE6 server, just tell management it's still tomcat, which it is.

          • 47. Re: Seam.Next discussion

            Hi, all!

            My experience with Seam 2 is very limited, I just worked some weeks in a project which used it. But then I was solicited to write a series of articles about what Seam 3 is, because a publisher received a lot of e-mails from confused Seam 2 users that just did not get it.

            As a developer that did not use Seam 2, I find the philosophy of Seam 3 attractive. I did not like the hugeness of Seam 2, which tends to be a problem with most Java frameworks. However, since I started to test Seam 3, I discovered it is just too buggy and confusing to be used at the same time it did not solve any specific problem.

            But I am the minor of your problems. I read a lot of opinions of Seam 2 developers and all of them are disappointed. Firstly, because Seam 3 is not Seam in any sense: it is just a reutilization of the trademark to some totally different kind of stuff. Also, this new stuff just does not do anything to help them in the way Seam 2 helped them wonderfully. And the lots of references to frameworks developers in the docs just sent the wrong signal: nobody seems interested in using Seam 2 for creating frameworks, but instead to create applications.

            To me honest, the focus on framework development feels like you are just playing around instead of trying to build a product, because you are the only framework developers in this landscape. You feel that, once these little tools were useful to you, it is useful for someone else - it may even be true, but this someone else is not your loyal community of Seam 2 users but a small theoretical group of developers. And I can bet that these developers will not use none of this because it is just too annoying to deal with such a lot of bugs that can be easily found.

            So, you are alienating your original, application-developer users. You also did not impress any new application developer with such an instable codebase, and I am not sure some framework developer will buy your talk either.

            What I, as an ignorant external observer, would suggest: just upgrade Seam 2. CDI is really beautiful so it can be interesting to replace the core of Seam 2 with it, but the rest should remain. Actually, the best case scenario is the one where Seam 2 developers just can compile their Seam 2 applications against Seam 3. Some stuff should be obsolete (e.g. @In and @Out) but it should be deprecated, not baned. Instead of trying to build something that nobody know what is or for what is useful, just make Seam 2 better using CDI. Do not rewrite from scratch, specially if rewriting from scratch will force an awful lot of happy customers to rewrite stuff themselves.

            I did not talked for myself until now. Now comes my rant.

            I spent two months writing an article about Seam 3 when articles about, let us say, CDI just took me two weeks. The time was spent on researching, testing, bug reporting and even some patchs. Anyway, I got the article done, sent it to my editor and the gave me it back with some change requests. I was searching for one of the changes when I discover that Servlet and Catch is merged with Solder now, which will force me to restructure my entire article.

            I now it is my problem, but it is Seam's problem, too. People will be afraid of writing documentation, tutorials etc. that will be obsolete in one week. Worse, people will be afraid of writing applications that will be obsolete in one week. You can ignore anything I say here (I do not know better anyway...) but one is well advised to avoid such recent, unstable changes.

            Also, there are three things that I would happily use: the @Veto annotation, the Expressions bean and Catch. Unfortunately, the Expressions bean did not worked for me if I separate API from implementation. If I do not separate them, then the Expressions does not work on Tomcat, apparently. It is ok because you do not want people using Tomcat anyway, but it is frustrating nonetheless and I just gave up.

            Then came Catch, which is great... and merged with Solder now, so I am forced to include Solder in my app if I want Catch. The attractiveness of the philosophy of Seam 3 mentioned in the third paragraph just dimmed out a bit more.

            Don't get me wrong, I hope that Seam become a huge success. However, Seam 3 is in crisis, and I think a sincere, direct opinion is what you need and look for.

            Good luck, sincerely!

            • 48. Re: Seam.Next discussion

              Thank a lot Adam,

              FULLY agree with all what you have said.

              As explained, seam 3 is not the continuation of seam 2. We can't even say that seam 3 is a framework. seam 3 name is confusing ;-(

              I know some people are still working on seam 2. So why not porting seam 2 to java ee 6 standards... Seam 2 users will certainly appreciate.

              At this time I still using seam 2. Once I have to move to java ee 6, I'll probably go away from seam 3 (and seam at the same time)...



              • 49. Re: Seam.Next discussion

                I know some people are still working on seam 2. So why not porting seam 2 to java ee 6 standards... Seam 2 users will certainly appreciate.

                The problem with this idea is that most of the features in the core of Seam 2 already exist in CDI/JSF2 Facelets, etc.  So while it might make sense to make a JSF2 version of Seam 2 to ease migration, I'm not even sure what a CDI version of Seam 2 would look like.  Most likely, it would end up being a bunch of CDI extensions and components.

                Sounds like Seam 3, doesn't it?   Maybe the Seam dev team isn't totally dumb after all. ;)

                My team works on a pretty big Seam 2 app, and I've been scouting ahead with CDI / Seam 3.   While there are a few things missing at this point, most of the features of Seam 2 are available in CDI/JSF2/Facelets2/Seam 3.   What about this is not a framework?   Is it not a framework because it's modular?   That's not a good argument.

                • 50. Re: Seam.Next discussion

                  @Adam - Sounds like this could be solved with some good documentation.   As you said, the Seam 2 core is CDI, and that's well documented.

                  I think you should give Solder a fair shake.   I've found it to be very useful.

                  Many of the other modules are very light on documentation and examples.   Maybe you could help out with that. :)

                  • 51. Re: Seam.Next discussion

                    Gaël Beaudoin wrote on Oct 03, 2011 03:46:

                    We're a very small company, and seam 2 brought us prodictivity, fast learning curve and powerful tools : RichFaces, Hibernate Search integration, PDF generation, easy email templating, fast / reliable CRUD operation (Home/Query), xxxx.page.xml metadata files are very useful to set security, description, page flow, parameters, RESTEasy integration, etc.


                    Seam 3 should be an integrated, fast and east to use framework.

                    My boss was impressed. I need xls export : 2 hours later, done. A PDF export here would be cool. 1 hour later, I'd have a first version. A REST API to use by mobile devices ? All needed stuff is here, examples are available and getting started is easy. I've never lost too much time figuring things out with seam2. It's like an almost endless box of tools and building box to play with.

                    Pretty much exactly my experience.  I think what we need is an all-in-one package, plus some 'integrated' documentation and examples.  I've been working with Seam 3 for a little while now, and it is definitely much slower going because I have to know where to look for the documentation of various features.  Every time I think how do I ... then I have to look around at CDI/JSF2/Facelets2/Solder/Security, or whatever.

                    You are exactly right about Seam 2 having comprehensive, end-to-end documentation that is very useful for application developers.

                    • 52. Re: Seam.Next discussion

                      Joshua Davis wrote on Oct 16, 2011 12:49:

                      Is it not a framework because it's modular?   That's not a good argument.

                      too much modularity generates complexity. what I first expect from a framework is an increase in productivity by providing me with needed modules, scripts ... out of the box, so I can focus on my business logic. something like seam2;-) all-in-one package is indeed a good option in this regards. Having to look for modules, versions ... we will probably be more productive with javaee6 and a good IDE.


                      • 53. Re: Seam.Next discussion

                        I see that all developers are complaining about some features that are not available in seam 3 like :

                          -- xls export,PDF export,JCR, REST API, Drools, jbpm, ... --- endless box of tools. (wide supply for module integrations with other frameworks/libraries like seam2)

                          -- Complete documentations and samples (The most important one!! seam 2 was very good in docs and examples)

                          -- Migration Plan from seam2 to seam3 and jee6, I see most developers are confused about seam3 vs seam2 and it's features that are included in JEE 6 now.

                        • 54. Re: Seam.Next discussion

                          I am migrating my seam2 application to seam3, many problems I encountered when used Seam3 modules.

                          I dislike the idea before I saw in the blog post(let the different communities hand over the seam 3 modules).

                          I think categorying the current modules into two parts is better:

                          1. Core modules(maybe some equivalent to solder in seam3 3.1.0.Beta3), provided seamless integration with standard JEE specs, such as

                          jsf(life cycle events, transaction integration, conversation management, basic extension, such as entityConverter, dataTable selection, simple file upload, graphicImage, resources, download etc),
                          el(I hope jboss.org continue support the jboss el in future), 
                          conversation(support none jsf web interface),
                          rest , 
                          bean valiation, 

                          2. umbrella projects/modules, provided by jboss.org or contributed by other developer. such as

                          tapestry5(there is a project provided extension for seam2, but I can not found such one for seam3),
                          jsf extension(pdf, excel, rss ui components),  
                          page flow, 
                          jbpm 3 task(is there some alternatives for this???), 
                          javafx2(I think there is module like this), 
                          payment(I hope there is a module for processing payment gateway), 
                          social(extended the security module, provided capability for accessing popular SNS), nosql(support popular NoSQl database access, CDI event based, transaction support if possible) 

                          The core modules use the same version, thus we can tell someone which version of seam3 I used, and the umbrella depends on the core modules and can use different version.

                          • 55. Re: Seam.Next discussion

                            There is problem with documentation, I agree. Seam had integrated documentation and it was good. Code snippets were based on full Seam.

                            In Seam 3 you have Seam doc's and Weld doc's. And after moving some modules to another projects you will have to look there for help. But it's not so big problem, some more tabs in firefox :) But the problem is with code snippets (and this problem is even in Seam documentation now): they are so basic that show only features of this one module, they doesn't show integration between modules.

                            Let's get for example Catch module - it's use is described only in one place. None of the other modules documentation use it to show integration. So if I try to use it, some exception are catched as in docs, and sometimes it simply doesn't work. I try to forward from such situation to for example error page and sometimes it forwards and sometimes not. I will not find example of whole application which uses Catch to catch all exceptions, so I don't know if it's lack of doc, or my fault that it doesn't work.

                            Seam team is wise and skillable, but they didn't write any final product based on Seam 3 so they don't understand us, developers. They tried to write forum and blog in Seam 2 and the only what we I say - it usually works. Forums are awfully slow and logs out quickly forwarding you to Seam homepage when you login. Blog times out during writing longer comment. Composing new message window uses some not widely used characters to format text, it doesn't allow for some commonly used chars.

                            So if core developers has written only some faulty Seam 2 application and none Seam 3 application - they don't understand us.

                            • 56. Re: Seam.Next discussion

                              We certainly hear you, and we're sorry we've let you down in this area. We're working a couple of fully integrated examples, and also better docs. In fact, we're planning on doing a 3.1.1 which will solely focus on documentation and better examples.

                              • 57. Re: Seam.Next discussion

                                That's really nice to know. Thank you for the update!

                                • 58. Re: Seam.Next discussion

                                  Jason Porter wrote on Oct 24, 2011 22:23:

                                  We certainly hear you, and we're sorry we've let you down in this area. We're working a couple of fully integrated examples, and also better docs. In fact, we're planning on doing a 3.1.1 which will solely focus on documentation and better examples.

                                  Good news!!!

                                  • 59. Re: Seam.Next discussion

                                    i researched all the current frameworks years ago and the first version of Seams i downloaded was 1.1. I developed on 1.2 for a year or so and started my move to Seam 2. I was a beginner then. I could jump right in with examples and seam-gen. I could get a project laid out in a matter of minutes after putting together my DB layer. That was 5 years ago. I started looking at Seam 3 when it was first leaked to the community.

                                    MY GOD WHAT DID YOU DO? I understand the concept but this release killed the buzz around this framework. And like others have already said, its no longer a packaged framework. Its a pile of tools. Though very powerful, it is entirely too complicated even for experienced guys that have been working in this since 1.1. I know Seam 2 inside and out. I can put together a site in a very short amount of time and pack a lot of functionality that would have taking me years with other frameworks. And though you guys have given me all this ability,,, Seam 3 took it away. I've tried several times to take a project forward only to be met with all of the problems mentioned above and more.

                                    I fear for the future of the project because of inability of new people researching it to even navigate through an example successfully.

                                    Now with all the reading ive done... and all the negative feedback i just gave.. i was leading up to this. I don't disagree with the direction. I think you guys are actually right in theory. I just don't know that the community can handle these birthing pains. You have to make this easier for the developers. We dont get paid to read and research. We get paid for delivering working software. Time matters.

                                    No joke. Gavin is a genius. And I'm sure I'm not the only one doing this,, but I have to sit back and wonder this time about Ceylon. Can he really pull this one off? So here is the long and skinny of it. I will continue to develop in Seam 2 because what do we really gain from moving forward? Do we have noSql and big data made easy for us? nope. I want a wrapper for entityManager that allows me to persist straight to something big. I want to use my old code as much as possible. And since the project is continually growing,, as is my code base,, i really dont need more complexity. So unless you can offer big data integration in Seam 3 with a nice little example of how everything works.  What have we gained through this gigantic restructuring, if the old way is faster application development. No real sell here, because if Gavin pulls off the new language,, here comes the next learning curve and all my existing code base is worthless. I'm pulling for him. I see the necessity of what he is doing. But the truth is, this project really felt him change his mind about the direction of the industry.