Learn how easy it is to sync an existing GitHub or Google Code repo to a SourceForge project! See Demo

Close

#265 classes in interleave mode and cardinality membership.

AMBER
closed-rejected
5
2012-06-05
2010-11-25
Bertrand Gaiffe
No

In content model of elements, we may refer to model classes in alternation, sequenceOptional, sequenceOptionalRepeatable, seuanceRepeatable and sequence modes. Why not interleave ? Moreover, why not being able to add an element in a classe with a min cardinality and a max one, i.e. <classes><memberOf key="someClass" min="something" max="sometingElse"/>. Of course, this makes sense only for sequence, alternation and interleave.

Discussion

  •  
    Attachments
  • Should also modify odd2odd.xsl so that the min and max attributes are not lost in "change" mode. Will do that when possible. Moreover, looks strange but trang seems unable to produce W3C shemas with cardinalities in the case of min and max (generates a simple choice ????).

     
  • I think the attached odd2odd.xsl is OK but did not test enough.

     
  •  
    Attachments
  • Lou Burnard
    Lou Burnard
    2011-02-22

    • milestone: --> AMBER
    • labels: --> TEI: Suggestion for ODD
     
  • By interleave, I mean if elements A B and C are members of class M, then : putting <ref name="M_interleave"/> yields <interleave>
    <ref name="A"/>
    <ref name="B"/>
    <ref name="C"/>
    </interleave>. That is, A B and C must be present in any order. I used that for metada components in which elements had to be there (with the "good" cardinalities) but wanted to relax somehow the order.
    Rq : in practice, you have to produce :
    <interleave>
    <group>
    <ref name="A"/> {n, m}
    </group>
    <group>
    <ref name="B"/> {p, q}
    <group>
    <ref name="C"/> {r, s}
    </group>
    </interleave>

    if you mix the cardinalities with this. That is you get any order provided the A's are grouped together, the B's are also grouped and the C's as well (this seems a relaxNG limitation). Therefore : BBB AAA CC is OK but ABACAB is not.

     
  • Laurent suggest that this ticket is divided in two.
    cardinaly restrictions = 3290834.
    interleave is probably not clear enough yet...

     
  • Lou Burnard
    Lou Burnard
    2011-11-08

    Martin asks whether this works in XSD? This ticket should be closed, and a new one on the interleave issue should now be posted.

     
  • Lou Burnard
    Lou Burnard
    2011-11-08

    • assigned_to: nobody --> rahtz
    • status: open --> pending
     
  • Piotr Banski
    Piotr Banski
    2011-11-08

    • assigned_to: rahtz --> nobody
    • status: pending --> open
     
  • Piotr Banski
    Piotr Banski
    2011-11-08

    Isn't it the other way round: XSD allows for stating min/maxOccurs, RNG does not have a straightforward mechanism for that. (?)

     
  • Lou Burnard
    Lou Burnard
    2012-04-17

    • assigned_to: nobody --> rahtz
    • status: open --> open-rejected
     
  • Lou Burnard
    Lou Burnard
    2012-04-17

    Agreed that since we cannot see any way of implementing this in DTD or XSD the proposal should be put aside. For discussion as part of the future of ODD road map.

     
  • Lou Burnard
    Lou Burnard
    2012-06-05

    Closed after further review at Council meeting in Ann Arbor

     
  • Lou Burnard
    Lou Burnard
    2012-06-05

    • status: open-rejected --> closed-rejected