The CXF project is highly dependent on XmlSchema. So much so that we, collectively, and I, individually, are willing to apply some significant effort to it.
While we run into point issues from time to time, we would really like to see a version of XmlSchema that has some significant overall modifications that are not really suitable to patches submitted on JIRA entries. Some of these are: 1) Move to Java 5 and use Generics to reduce the cast burden. 2) Improve performance. 3) Make the API more friendly to building schema from scratch. It's not like it's impossible, but there are a number of areas involving schema collections where the current situation leads us to slow and clumsy code. 4) Move the class model closer to the abstract data model of XmlSchema and further away from just representing the surface form of the XML representation. In particular, make it harder to create invalid schemata from the API. Just for experimental purposes, I made a branch of the current XmlSchema trunk in the CXF sandbox (https://svn.apache.org/repos/asf/cxf/sandbox/benson/xmlschema/XmlSchema) and did a first pass at the first half of 'step 1' above. Any readers of this can co that tree and look around at my work and criticize it. (At CXF, we allow any committer from any other project to commit in our sandbox, which aids us in inter-project integrations.) I would be happy to move this effort to a similiar sandbox in the XmlSchema universe. I would be happy to wake up and discover myself a committer over here, as well, and just do this work in a branch for the purpose. That would allow this work to be just another fully visible activity of WSCOMMONS, and perhaps that would attract some other contributors. While campaigning for committer-hood is generally bad form, I would submit for your consideration: WSCOMMONS-369 (pending patch) WSCOMMONS-272 (applied patch) WSCOMMONS-270 (applied patch) and some ancient discussion on the list, I think.
