On 15-02-03 03:19 AM, Piotr Bański wrote:
> Dear Martin,
> Thank you for enlightening me on the DOI issue -- now I understand your
> original point. As for:
> On 03/02/15 11:45, Martin de la Iglesia wrote:
>> This might work, but maybe that's beside the point: regardless of how
>> the <idno> content is defined in the ODD, I think the Guidelines
>> definition of @xml:base is unclear about whether it is applicable to
>> element content or not, or where exactly these "relative URI references"
>> are to be found. Wouldn't it be desirable to revise this wording into
>> something more precise?
> I think there are two issues here. The first is that the Guidelines
> should not even try to "define" xml:base, because they have no authority
> over this system -- they can at most quote the official definition and
> possibly comment on it (one comment may be that the original seems to
> have a historical, apparently DTD-data-type-inventory-related,
> pro-attribute bias).
Exactly. The TEI (or of course the customizing user) is responsible for
defining what is a URI and what isn't:
"This specification does not attempt to specify which text strings in a
document are to be interpreted as URIs. That is the responsibility of
each XML vocabulary. The question addressed by this specification is:
given a relative URI in an XML document, what base URI is it resolved
I believe that if you were to define the content model of <idno> to be a
URI, then you should expect a schema-aware XML processor to expand that
URI appropriately based on the @xml:base attribute on the element (since
the element content is a text node which is its descendant). But like
Piotr, I'd be intrigued to know if that actually works with processors.
> Another issue (and I think this is the spirit of your message, with
> which I agree completely) is definitely within our reach: the precise
> wording of the relevant section(s) of the Guidelines, to clearly report
> on what I mention above, among other things such as code examples and
> use cases.
> Best regards,