OK. I've been assuming that (regardless of the attr type) the scoping requirement is to allow for nested or subsequent topics in a single file that may have children with non-unique ids. In a ditamap, you can't nest a <map> element, so there is only ever one "limiting" id, thus no need for scoping.
But I'll post this question on the dita-users board. -----Original Message----- From: Hussein Shafie [mailto:[email protected]] Sent: Monday, August 14, 2006 2:27 AM To: Mark Fletcher Cc: xmleditor-support at xmlmind.com Subject: Re: [XXE] dita_map.css shows @href of source topicref even when a @conref is specified Mark Fletcher wrote: > I think XXE is wrongly not supporting an un-scoped conref in this case. > > The conref in question was in a map/topicref element. I originally > tried using a scoped href (file.ditamap#map_id/topicref_id), which XXE > successfully transcluded; but the toolkit threw an error, expecting a > non-scoped id (file.ditamap#topicref_id). So I think this is an XXE bug. We are happy when you find bugs in XXE, yes, really, but in this case, we are almost sure that the *only* *way* to refer to a <topicref> is "file.ditamap#map_id/topicref_id". * This because the @id attribute of a <topicref> has type "NMTOKEN" (not a real ID ==> needs to be scoped). * This in constrast with the fact that the @id attribute of a <map> has type "ID" (real ID). If we are wrong, that is, if "file.ditamap#topicref_id" is a correct URL, then DITA is *inconsistent*. If this happens to be true, we are not sure that we'll fix this problem in XXE.
