On 03/23/2011 08:52 PM, Tyrin Avery wrote: > > I did see that section of the spec. When I read the phrase: > > "For each vocabulary module in the referenced document, the referencing > document qualifies the common module with a subset of the constraints in the > referenced document." > > I assumed that this should mean that between two documents any vocabulary > that the two have in common (i.e. the most restrictive constraint) is used. > That should mean that any vocabulary that concept and reference topics have > in common should be conref-able. (which would make sense). I thought that was > further borne out in the examples in the section, which never use topic type > as a qualifier, which they definitely _should_ if they intended such a > drastic change. > > Also, it says in the generalization and constraints section that a > generalization processor "Looks for the first vocabulary module that is both > present in the target document type and that has a subset of the constraints > in the document instance" which also implies what I said above. > > So I think you should follow your first instinct, which was that it made no > sense the other way and change your implementation. I will use the workaround > however, with thanks. >
OK. We don't understand how this should really work so we'll change our implementation in XMLmind XML Editor v4.9 (and also in XMLmind DITA Converter, where this case has always been reported as a warning and not as a fatal error). -- XMLmind XML Editor Support List [email protected] http://www.xmlmind.com/mailman/listinfo/xmleditor-support
