3 Replies Latest reply on Nov 25, 2009 11:28 AM by jamesjmp

    The way systems (Wiki) fail, will it change with Seam 3 and JSF 2.0

      Today, for the second time, I lost a post.


      There use to be a forum to discuss issues with this site... it no longer exists...


      But that is, of course, not relevant as a post in this forum...


      Except for one reason:


      It is an example of the way systems fail with JSF/Seam:


      I wrote medium sized a post (it took me around 15 minutes, it was an answer and it is hard to get the Wiki Text Markup correctly between the blockquotes).


      I click Save...


      And I am greeted with this error page.



      After that I clicked the back button of my browser... but it was useless, my message was gone.


      Now, this is typical of web pages, nothing new with that... except because I had a similar experience with a system developed where I work (by someone else), and it
      did not lost my data... what fancy technology did it use? raw JSP to produce XML files that are read using raw XMLHttpRequest... and JavaScript that only works in IE6.
      What? No fancy JSF? No fancy AJAX remoting framework? and still did not loose your data? How is that possible? You may ask...


      Well it turns it is really simple! The code that sends data in the system developed  by this coworker sends the data using javascript, so handling a failure in the server is as
      easy as writing a try{}catch{} that means that by default, with this very primitive way of doing AJAX his application is more robust than a JSF application like this Wiki...


      If that is not a really serious indicator that JSF has a serious abstraction leak... then... why did this site lost my post? what would it take to modify its source
      so that it is resilient against simple errors such as JBSEAM-4171? Remember I am not asking here for fancy Google Gears like functionality, all I want is a chance to push the submit button again...


      Is that too much to ask?