On 04/08/11 17:07, Tomaz Erjavec wrote:
> But it turns out that things like<msItemStruct class="#nrss_type_sermon
> #nrss_author_ecclesial"> are invalid because of the @class value.
Are you sure? Looking at the definition of this attribute, I see that it
is defined (like other such attributes) in the ODD as
> <datatype minOccurs="1" maxOccurs="unbounded">
> <ref xmlns="http://relaxng.org/ns/structure/1.0" name="data.pointer"/>
so its value *is* repeatable.
> - the example gives the @class without the #, which is dangerously
> misleading - for a while there we thought, well, data.code for some reason
> does not use #, so we here happily writing @class="nrss_type_sermon
> nrss_author_ecclesial", which does validate, but is probably nonsense.
> So, would this be an error or two in the Guidelines?
Yes, this is an error. There should be a # on that value!
> And a related question: we actually thought to link the @class values to a
> taxonomy, but reading the definition, I see that it is meant only for
> linking to textClass.
That is clearly wrong. Well spotted!
I think this is a fixable bug : obviously @class should point to one or
more <category> elements within a <taxonomy>, or to one or more
classification entries in some external thesaurus.
> But http://www.tei-c.org/release/doc/tei-p5-doc/en/html/ref-textClass.html
> defines it
> <textClass> (text classification) groups information which describes the
> nature or topic of a text in terms of a standard classification scheme,
> thesaurus, etc.
> I would then imagine that each text gets only one textClass; in which case,
> how could you have more than one value for @class, and what is the point of
> having @class anyway, given that all msDesc belong to the same<TEI>? (well,
> except for the borderline case where you would have a teiCorpus of
> Any illumination would be much appreciated!