Thanks, Hussein. I asked Don Day directly about the scoping issue and here's his reply. (He had been ignoring [XXE] mail and didn't see your cc to him on this subject.)
"Robert [Anderson] informs me that the unscoped form is actually correct, and that the conref code in DITA OT is properly handling the case. I asked him to respond to you and Hussein with the justification, which is partly explained in the DITA Architectural Spec." So XXE's interpretation was correct. To the subject of the format="ditamap" method of including map references: I understand you guys are swamped. But given the fact that you delivered conref support WAY earlier than I had ever hoped, I'll keep my fingers crossed that you might sneak this into 3.4 ;-) (It *is* frequently used and would probably be a differentiating feature between XXE and other editors out there. I've posted a question on the dita-users group to see if anyone else currently supports this.) -----Original Message----- From: Hussein Shafie [mailto:[email protected]] Sent: Monday, June 19, 2006 2:18 AM To: Mark Fletcher Cc: dond at us.ibm.com; xmleditor-support at xmlmind.com Subject: Re: [XXE] questions about maps and conrefs Mark Fletcher wrote: > > I'm taking a look a v3.3.0 and I'm trying to use a conref from a > topicgroup in map A to a topicrgroup in map B. Should this be supported? Sure. > I'm getting this error message: > > /map/topichead/topicgroup: > conref target "Untitled.ditamap#map_conref_test/group_conref_test" not > found > > In this path, Untitled.ditamap is in the same directory as the > referencing map; map_conref_test is the map @id and group_conref_test > is the topicgroup @id. Is there something wrong with my path syntax? I'm really not a DITA expert but I would say that the URL should be "Untitled.ditamap#group_conref_test". The id attribute of a topicgroup (unlike the id of a topicref) has type ID and, in my understanding, should not be specified as a *scoped* id (i.e. map_conref_test/group_conref_test). I have reproduced what you did (see attached screen shot) but instead of typing paths (which is tedious and error prone), I have used the "F2 c"/"F2 v" keyboard bindings ("F2 c"=copy conref target, "F2 v"=paste to conref source) as explained in http://www.xmlmind.com/xmleditor/_distrib/doc/dita/setconref.html, and everything worked smoothly. Now, we may be 100% wrong in thinking that an id attribute having type ID should not be scoped, and in such case, it would be nice to tell us where all this is clearly explained[*] and we will fix it in next release. Note that there should be no problem at all with topics, tasks, concepts, etc. > A related question: as you mention in your doc, using @conref in maps > will not be commonly used. (My @conref test above is just that--a > test.) However, I do use (and I think many others will use) the > @format="ditamap" option in <topicref> to include other entire maps in > publishing output. In this case, during publishing, the topicref is > replaced by the contents of the map it points to. The DITA-OT fully > supports this, so my publishing is working just fine. But I'm > wondering if you would be able to support this as yet another > inclusionProcessor type. Currently, I'm using a whole bunch of CSS > xpath statements to show a "read-only" view of the first-level map > children. But it would be very cool if you could actually transclude this content. We didn't know about this feature (which does not leverage the conref mechanism). Once again we are not DITA experts and it will take us years to learn which are the best practices of DITA authors, and therefore it will take us years to implement these best practices. We are not going to implement what you have requested in next release because we plan to ``integrate'' the DITA-OT (don't ask us how!), and we think that's a sufficient step for v3.4. --- [*] In order to learn about the fine points of conref, we had to read the XSLT source code of conref.xsl, which is part of the DITA-OT.
