[topicmapmail] <mergeMap> questions

[Jan Algermissen]

Dear list members,

I have two questions about processing the <mergeMap> element
correctly. Does anyone of you have a reasonable answer to them ?



1) Suppose that an XTM processor has already parsed a referenced topic map
   WITHOUT added themes (due to processing a <topicRef> element) and later
   processes a <mergeMap> element that references the same map WITH added themes
   on it. How should the processor handle this situation?

   a) Process the map again, doubling the associations of the  map, this
      time WITH the added themes.

   b) Process the map again, adjusting the scope of the associations rather
      than doubling them.

   c) Don't process the map a second time, thus igonring the added themes ??

   d) Report an error.


Example code:


<topicMap>
  <topic id="t1">
    <baseName>
      <scope>
        <topicRef xlink:href="www.topicmaps.org/xtm/1.0/language.xtm#en" />
      </scope>
      <baseNameString>The english name of t1</baseNameString>
    </baseName>
  </topic>

  <!-- .. -->

  <mergeMap xlink:href="www.topicmaps.org/xtm/1.0/language.xtm#en">
    <subjectIndicatorRef
      xlink:href="www.topicmaps.org/xtm/1.0/language.xtm#xtm1.0-psi-language" />
  </mergeMap>
</topicMap>

I think that the result should be the same  for any ordering of the XTM elements,
right?


2) Do added themes accumulate when recursivley processing <mergeMap> elements ?
   Suppose an XTM processor is parsing a <mergeMap> element with the added themes
   t1 and t2 and the referenced map contains another <mergeMap> element with added
   themes t3 and t4. What are the added themes on the third map, t1,t2,t3 and t4
   or just t3 and t4 ??


I am unable to find a sound solution to these two questions, and just in case that
there is no *single* answer, it would be interesting to see what the general
understanding (=best solution ??) of these issues is.



Jan