Blogbody

I found, via TSS, possibly the nerdiest quote ever. Any respect I may have had for Prevayler (in which I admittedly know nothing about and yet still claim it is a pipe-dream), died when I read this :

As RAM becomes cheaper, I believe that the progressive adoption of prevalence is unavoidable. There are great barriers of lack of knowledge and fear to overcome, but until there, the technologies you mentioned will still be fixed inside a retrograde manager's mind (like COBOL). We are babies born inside the Matrix: brutely degenerated by the use of databases. To start to use prevalence gives a sensation similar to Neo's: "Why do my eyes hurt?" Morpheus answers: "You've never used them before."

Posted by Patrick Lightbody at 12:03 PM | Permalink | Comments (0)

I added, as a prototype only so far, XmlHttpRequest-based client-side validation in WebWork 2.1.7.

Right now it is set up such that on every "onblur" event, a request is sent out to a special ValidationServlet, which sends back an XML response indicating the validation results. Then, using DOM, we insert the response in to the form dynamically.

This is just the start of some of the things I want to start doing with WebWork, all in effort to make the UI part of the framework much more rich.

Posted by Patrick Lightbody at 07:54 AM | Permalink | Comments (2)

Mark Roth recently asked for feedback on the Faces 1.2 and JSP 2.1 specs (via a chat mid-work day, unfortunately).

My response is this:

Mark,
I'm probably not going to be able to make it to the chat, but I did want to raise a one issue about the new JSP spec: I'm worried that it is becoming too married to JSF. It already tied itself too tightly to JSTL by stealing the "${}" syntax, and now I hear it might also take "#{}". I strongly encourage any future JSP spec to allow for the tag author -- not just the page creator (JSP editor) or web deployer (web.xml editor) -- to be able to specify if they want their tags to have attributes parsed for ${} or #{}.

As a developer for WebWork, it disturbs me to see keywords suddenly unavailable for JSP tags. We had to settle on %{} because ${} was taken out of our control. Fortunately, we could work around this. If #{} is used in a similar manner, the problems will be much bigger. OGNL uses #{} to represent maps and using this syntax would really hurt WebWork and Tapestry as they both use OGNL.

As an aside, I'd like to see JSP and JSF specs encourage more feedback from communities other than Struts. The Tapestry, Spring, and WebWork communities have a lot to offer and I haven't seen too much reaching out to them (or at least to me -- but maybe that's my fault).

Posted by Patrick Lightbody at 10:55 AM | Permalink | Comments (2)