coming late to this party…
> On 26 Mar 2015, at 21:02, Stuart A. Yeates <[log in to unmask]> wrote:
> I was reusing an EEBO TEI file as a template for another context and I
> noticed a couple of unusual features. I'm not sure whether these are
> design features or accidents of history
> My working example (picked at random) is
> *) The primary identifier within the XML-based system is A00005, but
> there is no xml:id="A00005" or similar easily XSLT-matchable attribute
> (I was expecting to see it on the root element).
Design. it is there in an <idno> element.
> *) There is a complete lack of xml:id / id tags
Design. unless something needs pointing to, it has no ID here
> *) There is no xml:lang tag covering the textual content in the header
Accident, possibly. Not sure.
> *) The <availability/> tag has a license URL, but doesn't use the
> <licence/> tag as the container for it. This is a significant barrier
> to automated license detection and thus automated reuse.
Accident. I will try to change these globally, as I added that stanza
> *) The <language> tag and xml:lang both use three letter language
> codes (I believe that http://tools.ietf.org/html/bcp47 forces us to
> prefer two letter language codes where available)
Accident. Would you like to check if they can all be changed losslessly
and if so push up a lot of changes?
> *) The textual content of the header contains a LOT of proper nouns
> with corresponding websites, but none are linked using <name/> or
Design. No-one ever claimed the texts are marked up to that degree.
Chief Data Architect
University of Oxford IT Services
13 Banbury Road, Oxford OX2 6NN. Phone +44 1865 283431