Paul Rabin wrote: > > At 10:11 AM 7/25/01 -0400, Jonathan Borden wrote: > > >Ok I understand that the attribute has a different name (that is one > >difference). I still don't understand how it functions in a different way. > >In specific an implementation of XInclude acts _as if_ the included document > >was in the including document. Copying/reparsing etc are merely > >implementation issues. What I mean by saying that they are equivalent is > >that the "Infoset" provided by both mechanisms should be the same, if not > >please explain how the stored XML Infoset is different. > > If there are three documents, a.xml, b.xml, and c.xml, and both a.xml and > b.xml contain links to c.xml, then > > (1) both a.xml and b.xml appear to contain c.xml > (2) there is only one copy of the c.xml infoset items > (3) any change to c.xml infoset items made via a.xml, b.xml, or c.xml is > visible to the others.
This is a perfectly reasonable question. The XInclude WD specifically distinguishes XInclude from External Entities based on Infoset considerations: "Processing of external entities (as with the rest of DTDs) occurs at parse time. XInclude operates on information sets and thus is orthogonal to parsing. " > > Is this is how you would expect an XInclude implementation to behave? yes. > When > our link mechanism was implemented, early versions of the XLink > specification were available, but the XInclude specification didn't exist > yet, and so of course it wasn't a consideration. sure. now that it _is_ at least a WD, we should try to converge. -Jonathan ---------------------------------------------------------------------- Post a message: mailto:[EMAIL PROTECTED] Unsubscribe: mailto:[EMAIL PROTECTED] Contact adminstrator: mailto:[EMAIL PROTECTED] Read archived messages: http://archive.xmldb.org/ ----------------------------------------------------------------------
