Couple of things. The plan was to do something quick to sort out the data model we needed internally. The loader uses raw DOM for now but should migrate to something better (XMLBeans) in the future; this will probably happen in conjunction with the JSR88 tool impl that needs to write out valid documents.
For a more general solution, we have been thinking about persistence to XML in general (e.g. in a form that the CMP impl could use). This hasn't got much beyond the "neat idea" stage, but if anyone want to pick it up the wiki may be good place to start. -- Jeremy > -----Original Message----- > From: Jacek Laskowski [mailto:[EMAIL PROTECTED] > Sent: Thursday, October 02, 2003 3:54 AM > To: [EMAIL PROTECTED] > Subject: Re: How to work with XMLs? > > > James Strachan wrote: > > > The xbeans module was an experiment and used BEA's XMLBeans (now at > > Apache), JAXB and Castor. Out of the 3 the best solution > appeared to be > > XMLBeans. > > Though we've decided to write our own beans by hand & hand > code the > > marshalling ourselves via Digester (I think) > > Digester? Well, that would help a lot, but alas Geronimo > doesn't use it > at all. I spent some time yesterday trying to figure out how > it's done > and it looks to me that it's necessary to handle binding elements and > attributes by ourselves. It means we don't use any tool > (besides Xerces) > to work with XMLs. That's different than how I used to manage > it, but it > gave me another idea. > > What about a service that would allow handle XML binding > automatically? > Whatever it means. What benefits would it provide besides > being opaque > to the 3 listed above - I don't know. I can't imagine how > it'd work at > all, but hope to see others who pick it up and materialize :-) > > -Jacek > >
