Radu's question and analysis of subjectScheme content in the DITA 1.3 spec

[Kristen James Eberlein <kris@eberleinconsulting.com>]

I've read through both the Language Reference and ArchSpec topics carefully (and found a few glitches, which I've sent to the list).

I think we clearly and consistently make the following points:

• For bound enumerations, the category value is NOT included in the list of permitted values; it just serves as the reference to the children values. We repeatedly explicitly state this in examples.
  
• Key references to controlled values are resolved to key definitions "using the same precedence rules as apply to any other key." http://docs.oasis-open.org/dita/dita/v1.3/os/part1-base/archSpec/base/subjectSchema.html#subjectSchema
  

Given Radu's example:

!DOCTYPE subjectScheme PUBLIC "-//OASIS//DTD DITA Subject Scheme Map//EN" "subjectScheme.dtd">
<subjectScheme>
    <subjectdef keys="floatLeft">
        <subjectdef keys="left"/>
    </subjectdef>
    <subjectdef keys="outputclassFig">
        <subjectdef keys="none"/>
        <subjectdef keyref="floatLeft"/>
    </subjectdef>
    <enumerationdef>
        <attributedef name="outputclass"/>
        <subjectdef keyref="outputclassFig"/>
    </enumerationdef>
</subjectScheme>

I think processors need to resolve the key reference to "floatLeft" in place and so interpret the bound set of controlled values as follows:

    <subjectdef keys="outputclassFig">
        <subjectdef keys="none"/>
        <subjectdef keys="floatLeft">
          <subjectdef keys="left"/>
        </subjectdef>
    </subjectdef>

Note that the category value "outputclassFig" is not a permitted values for @outputclass.

Should we add this as an example to the DITA 1.3 errata?

--
Best,
Kris

Kristen James Eberlein
Chair, OASIS DITA Technical Committee
Principal consultant, Eberlein Consulting
www.eberleinconsulting.com
+1 919 682-2290; kriseberlein (skype)