[tei-council] conformance draft

[James Cummings]

Conal Tuohy wrote:
> I have read JC's "conformance" draft and I like it very much. Thanks James!

You're welcome, but I'm sure not everyone will be so positive. :-)  I know for a
fact that a number of people don't like this moral high ground of namespaces.

> I have one comment there about the use of namespaces in "renaming" schemas.
> There is a kind of "loophole" there which I think should be tightened.

I think I understand the loophole to which you refer, but might wish to tighten
it in a different way.

> If I read it correctly it seems to imply that Renaming Subset Schemas should
> have a special exemption to allow them to define new names in the TEI
> namespace, if it does so using <equiv> to relate them to existing elements.

I'd say that these aren't 'new' elements, they are simply 'renamed' elements,
syntactic sugar if you like.

> <quote>TEI-Conformant documents must not extend the TEI Namespace with
> additional elements and attributes except by means of a TEI Renaming Subset
> Schema produced from a TEI ODD file which properly documents the renamed
> elements or attributes through the use of the <equiv> element.</quote> It
> seems to me that a schema which renames an existing TEI element should do so
> in a new namespace. The semantics of the renamed element are the same as
> before, of course, but I'm not sure that automatically implies that the new
> element should belong to the TEI namespace. The issue to me is whether we
> should allow anyone to add names to the TEI namespace, and the question of
> whether those names are just new names for old elements or new names for new
> elements is not really relevant. The relevant question, to my mind, is
> whether those names are defined by someone independently of the official TEI
> schema development process. My rule of thumb is: if multiple people are
> defining names in the same namespace, then they should be coordinated; if
> they are not coordinated, then they should use distinct namespaces, otherwise
> there is a risk of collision.

Ok, I understand that.  Perhaps we should be distinguishing here between local
processing documents and those for 'interchange'.  We could say that a document
which validates against a TEI Renaming Subset schema should only be interchanged
with others when the changes have been reversed?

> 
> e.g. if I rename tei:div1 to tei:book, and someone else renames tei:text to
> tei:book, then we have a name collision. Whereas if I rename tei:div1 to
> my:book, and someone else renames tei:text to their:book then there's no
> conflation (assuming that my: and their: are prefixes bound to distinct
> namespace URIs, of course).

I don't think there is an assumption that documents validating against different
TEI Renaming Subsets are in any way compatible or interoperable.  Perhaps we
should tighten up the language to make sure.  Documents validating against any
individual TEI Renaming Subset schema don't have this problem, do they?  This
only occurs when I try to combine your documents with my documents?  Conformance
shouldn't imply interoperability... but two people with documents that validate
against a TEI Renaming Subset *can* interchange documents as long as they
convert them back to TEI Pure Subset first.

> On a similar subject: shouldn't the official TEI translations of element
> names also use distinct namespaces? e.g. "http://www.tei-c.org/ns/1.0#es" for
> the Spanish translation?

I'll leave that one to Sebastian, he wot knows internationalisation.

-James

-- 
Dr James Cummings, Oxford Text Archive, University of Oxford
James dot Cummings at oucs dot ox dot ac dot uk