Martin Holmes wrote:
>I fall rather in the middle on this one. I understand exactly what Dan
>means, but I take the point that language is perhaps not the perfect
>analogue. Generically, this is the kind of thing we're talking about:
> <descriptionOfThing xml:id="ThingID">
That's nothing like a feature structure then. It says that "ThingID" has
two properties, but the properties only have content, and are not
distinguished. Does it mean that the unnamed property is both "blah" and
"stuff" or what?
> <div associateWith="ThingID">
A mechanism like this already exists. It is called "decls". Elements
which are members of the "declarable" class can be pointed to by any
member of the "declaring" class (using the attribute "decls").
>The pattern is that something is described in the header (maybe by means
>of a feature description, maybe using other tags such as creatorApp, but
>with a unique ID) and then elements elsewhere in the document are
>associated with it by means of an attribute pointing to the ID value.
>The pattern is useful for all sorts of different types of item, from
>linguistic feature structures to creatorApp. I think Dan's actually
>using the term "feature structure" in a general sense, rather than
>referring to fsd and fs elements.
"feature structure" is a technical term with a specific sense both in
linguistics generally and in the TEI in particular. I dont find it
helpful to misuse it in this way.