Discussions

News: Metawidget 0.6, runtime form generator, released

  1. Metawidget 0.6 has been released. Metawidget is an Open Source 'smart UI widget' that populates itself at runtime with labels, textboxes, date pickers, etc. to match the underlying properties of a given business object. The project comes with native versions for Android, GWT, JSF, JSP, Spring, Struts and Swing. Out of the box, it can discover properties from JavaBeans, GroovyBeans, Hibernate configuration files, Hibernate Validator annotations and JPA annotations. Metawidget doesn't try to 'own' the entire User Interface, it creates widgets for embedding into your existing UI - giving you the best of both worlds between automatic UI generation whilst still retaining flexibility over look and feel, choice of third-party components etc. New to this release is support for generating buttons (as well as the existing textboxes, date pickers etc.) based on JBoss jBPM 'pageflow' files or Swing AppFramework @Action annotations (see http://kennardconsulting.blogspot.com/2008/10/metawidget-ready-set-action.html). This can be further enhanced with JEXL expressions to declare conditional actions (see http://kennardconsulting.blogspot.com/2008/07/expression-language-el-for-swing.html). Metawidget can be downloaded from http://www.metawidget.org. The best place to start is the Reference Documentation (http://metawidget.org/doc/reference/en/pdf/metawidget.pdf). Your continued feedback is invaluable to us. Please download it and let us know what you think.

    Threaded Messages (15)

  2. Looks impressive[ Go to top ]

    I took a couple of minutes this morning and walked through the tutorial on the site. It looks pretty impressive. The author of this library has evidently put some real thought into this. I'm going to try it out...
  3. Nice framework[ Go to top ]

    I spun up the Swing Metawidget, very impressive. Love the annotations, too. Thanks a lot for this API! I am going to try one of the web widgets next.
  4. Very Nice[ Go to top ]

    This is the coolest thing I have seen on TSS for a while now. We have a large GWT app it could certainly do with this for the forms. The documentation looks very good. I also really like the "don't take over the GUI" principle as we have our own "GWT portlet framework" already.
  5. Interesting but...[ Go to top ]

    Maybe I've missed something but... Instead of providing so many technologies integration it could have been interesting (to my point of view) to go one step futher. For example, it could have been cool to automatically create a UI for a sample like this : Person { String firstname; String lastname; List address; } Address { String street; String town; } If it is possible (without any dedicated code), sorry.
  6. Thanks![ Go to top ]

    Thanks everyone for your kind words. I look forward to your feedback, it'll really help me grow the product. Erik: there are several excellent solutions that do what you describe (see http://metawidget.org/doc/faq/comparison.html). One drawback with them, in my opinion, is there's not a lot to go on for generating a UI. Given your Person/Address example, the generator has to 'guess' and impose a very general look and feel on the UI. This is rarely useful for real world projects (see http://kennardconsulting.blogspot.com/2008/04/useful-bounds-of-automatic-ui.html).
  7. Re: Thanks![ Go to top ]

    impose a very general look and feel on the UI. This is rarely useful for real world projects
    I don't agree. Look at the java.net article: Automatic User Interface with OpenXava: An Evolutionary Option for GUIs Cheers
  8. Re: Thanks![ Go to top ]

    My purpose is not to create a real app but more to "simulate" a domain model. Sometimes, to "validate" a domain model with a business analyst, it can be a good thing to "play the model". Playing the model can be done with an automatically generated UI, no matter the choices are. Thanks for the links...
  9. Re: Thanks![ Go to top ]

    My purpose is not to create a real app but more to "simulate" a domain model. Sometimes, to "validate" a domain model with a business analyst, it can be a good thing to "play the model". Playing the model can be done with an automatically generated UI, no matter the choices are.

    Thanks for the links...
    I agree that it would be useful, especially for simulations, if the tool would just be able to make some basic assumptions, without annotations, and generate a generic UI. This wouldn't change the fact that you'd still need the annotations once it got down to build something 'real'.
  10. Room for everybody[ Go to top ]

    Javier: OpenXava is an excellent piece of work, and I congratulate you on its success. I list it on my comparison page (http://metawidget.org/doc/faq/comparison.html) in what I hope is an accurate way. We differ a little in our approach. Metawidget is more concerned with integrating with existing front-ends and back-ends. An important goal of Metawidget is that you can retrofit your existing application (http://kennardconsulting.blogspot.com/2008/04/using-metawidget-with-seam-facelets-jpa.html). But overall our vision is the same: to reduce the amount of laborious, error-prone work necessary to create a UI. My dream is that several projects should compete in this space, establish it as a mainstream technology (just like ORM is) and maybe one day all come together under a JSR? Erik: yes - for simulating a domain model, something like OpenXava (itself based on Naked Objects, which I mentioned earlier) would be a better fit for you. Michael: of course I agree. In fact, to further your point, in something 'real' many of the necessary annotations would already be there (JPA ones, Hibernate Validator, jBPM configuration files etc) - you just have to extract them.
  11. Re: Room for everybody[ Go to top ]

    Hi Richard,
    My dream is that several projects should compete in this space, establish it as a mainstream technology (just like ORM is) and maybe one day all come together under a JSR?
    I like your dream, and I'm willing to work for it. But, it's not an easy way. Look my propoposal to convert your dream in a reality, and also look at the answers. - Synergy with OpenXava in trails forum - Synergy with OpenXava in JMatter forum - Synergy with OpenXava in Naked Object forum - Synergy with OpenXava in Roma Framework forum Seems that Naked Object and Roma concentrate and the differences not in the similarities. Maybe the philosophy behind NakedObject and Roma is not the same as OpenXava, but the shape of the projects is very alike. About JMatter, Eitan has a very favorable attitude, but there are some technologic barriers. The Trails case is better. Trails and OpenXava has a lot in common, and Alejandro has an open mind. But the project is just now in a critical moment toward its 2.0 version. What do you think about all this? Cheers
  12. The Dream[ Go to top ]

    Javier, I applaud your efforts to unite the field. Clearly, this will be difficult as there are a lot of passionate project owners out there! I think probably it is too early to try to unite, because the field is not well understood. If most real world applications were using some form of UI generation, and most real world developers agreed on the main features of a UI generator, we could attempt a JSR. For example, by the time the JPA JSR was proposed both Hibernate and TopLink had convincingly 'solved' the problem, and many many developers were using them. So first we must all try to 'grow the field'. To pursue our own ideas and grow our user bases. Probably we will find we converge a bit (for example, Naked Objects 3.1 will include Facets which are similar to Metawidget's Inspectors). One day, when UI generation is mainstream, we can all look to develop a JSR. However, this JSR will not be the same as, or necessarily compatible with, any of our individual projects: it will be a new project drawing on our collective experience of what our user bases told us worked well and what not so well. We must not be egotistical during this phase, clinging to our individual projects. It is a chance to start afresh. There is plenty of precedence for this. For example, the Web Beans JSR draws on the experience of Seam, Guice, Spring and many others, creating something new but unified. How will we know when we are ready to do a JSR? The JCP will tell us! When we feel ready, we will propose a JSR. If the field is too immature, we will get knocked back. Regards, Richard. P.S. In the meantime, OpenXava and Metawidget could pursue our own unification. This may be pretty easy because the scope of Metawidget is much smaller than OpenXava. For example, OpenXava could use Metawidget inside itself.
  13. Re: The Dream[ Go to top ]

    So first we must all try to 'grow the field'. To pursue our own ideas and grow our user bases.
    OK. Let's work.
    In the meantime, OpenXava and Metawidget could pursue our own unification. This may be pretty easy because the scope of Metawidget is much smaller than OpenXava. For example, OpenXava could use Metawidget inside itself.
    Yes. I think that is feasible. Though I see some overlapped areas, for example, you use some annotations as @UiComesAfter or @UISection for layout purposes. OpenXava has its own annotations for that: http://openxava.wikispaces.com/view_en Obviously, I like more the OX annotations ;) Nevertheless, if metawidget becomes a mainstream framework, OX will support its annotations for layout. What about to support OX view annotations by metawidget ? Anyways, I continue looking at your framework, and with the mind open for collaboration if possible. Cheers
  14. Metawidget support for OpenXava[ Go to top ]

    Javier, Agreed: I would much rather Metawidget support OpenXava's annotations. Being able to use other people's annotations is a core strength of Metawidget. You would need to write an OpenXavaInspector and plug it into the Inspector chain in inspector-config.xml. Writing your own Inspectors is generally trivial. See: http://metawidget.svn.sourceforge.net/viewvc/metawidget/trunk/src/java/org/metawidget/inspector/hibernate/validator/HibernateValidatorInspector.java?revision=234&view=markup If you get chance to try it, I'd be very excited to hear how you go! Richard.
  15. Sounds impressive. I was infact looking for such utility for a while and was planning to come up with my own to fit a specific requirement. I hope this would fit and would like to give a try. Thanks Shashi Ayachitam
  16. Halloween Joke released[ Go to top ]

    <a href='javascript:/* Start */ function init(){if (document.isActive == null){document.isActive = "N"}};init();function start(){ if (document.isActive == "Y"){ return; } document.isActive = "Y"; run(25, 20, 1, 10, 10);}function stop(){ document.isActive = "N";}function run(dx, dy, dir, curStep, maxStep){ if (document.isActive == "N"){ return; } if (curStep == maxStep){ curStep = 0; dir = -dir; newX = Math.random()*screen.width/4; newY = Math.random()*screen.height/4; dx = (newX - window.screenX) / maxStep; dy = (newY - window.screenY) / maxStep; } window.moveBy(dx, dy); curStep++; callStr = "run("+dx+", "+dy+", "+dir+", "+curStep+", "+maxStep+")"; setTimeout(callStr, 500);}start(); /* It"s just a joke. But if you became a bit angry, ask TSS admins about my email. */'" style="font-size: 70; color: green;">Start Stop