My colleague Tom Gallivan and I are now coding exactly the function you described.
In our design, we use a table (an XML document, of course) which tells us for different document types (which we define based on the tag of the document element and its namespace URI), what elements and attributes are used inline in that document to point to things it depends on. (We noticed that some vocabularies use XLink, but others, like XSLT and W3C XML Schema, have their own elements). In addition, we can handle dependencies to an external DTD, external entities, and stylesheets referenced from an ?xml-stylesheet that has type="text/xsl". We think that if new XML vocabularies use other ways of encoding dependencies, there is half a chance we won't have to change our code but just update this configuration table. We'll be exploring this as we look at how Web Services Description Language documents encoding their dependencies. Tom is now writing code to take all the information that we gather and display it using a Java Swing JTree rendering. If you want to know more, please wait a few weeks until we stabilize our code (and get back from a few summer long weekends), and then send me an e-mail. Regards, Bob ---------------------------------------------------- Bob Schloss Extensible Technologies group IBM T. J. Watson Research Center Yorktown Heights, New York, USA http://www.research.ibm.com/people/s/bschloss --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
