[tei-council] Two quick questions about Schematron constraints

[Lou Burnard]

On 15/06/12 19:46, Martin Holmes wrote:

> In this constraint (from delSpan):
>
>     <constraintSpec ident="spanTo" scheme="isoschematron">
>       <constraint>
>         <assert xmlns="http://purl.oclc.org/dsdl/schematron"
> test="@spanTo">The spanTo= attribute of<name/>  is required.</assert>
>       </constraint>
>     </constraintSpec>
>
> does<name/>  automatically get replaced by ancestor::elementSpec/@ident?

Judging from the output of the detest test file, yes. I think it's dark 
magic though.

> What happens in the case of constraints on attributes? There appears to
> be only one (on @targetLang in att.pointing), but if<name/>  were used
> there, would it be replaced with @targetLang, or with the ancestor's
> @ident (att.pointing)?

No idea, but my guess is neither : I think it would be replaced by the 
name of the gi of the element which was being processed when the rule 
fired. Try it!

>
> Secondly, these constraints appear on the root TEI element:
>
>     <constraintSpec ident="c1" scheme="isoschematron">
>       <constraint>
>         <ns xmlns="http://purl.oclc.org/dsdl/schematron" prefix="tei"
> uri="http://www.tei-c.org/ns/1.0"/>
>       </constraint>
>     </constraintSpec>
>     <constraintSpec ident="c2" scheme="isoschematron">
>       <constraint>
>         <ns xmlns="http://purl.oclc.org/dsdl/schematron" prefix="rng"
> uri="http://relaxng.org/ns/structure/1.0"/>
>       </constraint>
>     </constraintSpec>
>
> What are they actually doing, if anything?

I think this is more dark magic, the effect of which is to bung the 
required namespace declarations into the isosch file generated during 
the P5 build.