[tei-council] content model of <surface>

[Sebastian Rahtz]

I am mildly horrified to see that we have got ourselves a non-deterministic content
model in here at some point, which has not been checked for (but will from now on)

If you look at this:

<content>
    <rng:group>
      <rng:zeroOrMore>
        <rng:ref name="model.glossLike"/>
      </rng:zeroOrMore>
      <rng:zeroOrMore>
        <rng:ref name="model.graphicLike"/>
      </rng:zeroOrMore>
      <rng:zeroOrMore>
	<rng:choice>
	  <rng:ref name="model.global"/>
	  <rng:ref name="zone"/>
	  <rng:ref name="line"/>
	  <rng:ref name="surface"/>
	  <rng:ref name="surfaceGrp"/>
	</rng:choice>
      </rng:zeroOrMore>
    </rng:group>
  </content>


you will realize that if <certainty> occurs as the only child  <surface>, it is a member of model.global (via model.global.meta) and of
model.glossLike (via model.cerLike), and thus we don't know where it came from (them's the rules of determinism).

I'd be interested to hear of a nice clean solution to this.... 

I suppose someone does know why we want <surface> to allow children of (eg) <pb/>?

Sebastian