Bruce Robertson wrote:
> Thanks, James, for your prompt and helpful reply. My first concern, of
> course, was that I was looking at this sideways. (Incidentally, does
> anyone know of a schema visualization tool that allows one to inspect
> what the possible parent elements of a given element might be? My
> students might have found this information useful if it were given on
> the P5 documentation site, since they could easily find a pertinent
> element, but sometimes could not easily tell what enclosing elements
> were required to produce a valid document.)
I believe oXygen has a schema visualization tool in it, but I must admit
I've never really used this since I just consult the TEI reference
material. Isn't what you're asking for the 'Used by' section of the tei
reference pages? I.e. If we look at:
We can see that event is allowed inside: event model.persEventLike
model.placeEventLike ...clicking on those class links then gives you a
list of what elements are members of this class. Similar reference
documentation can be produced by Roma for a customised schema, where you
may have added something to these classes.
> I do have another couple of questions about using the P5 event tags as
> historical markup, though. One of the two places that <event>s can be
> subordinated to <persons> is in the header, under <particDesc>. This
> is rather narrowly defined thus: "describes the identifiable speakers,
> voices, or other participants in a linguistic interaction." Now, if
> the text were an historical narrative whose events we wished to
> encode, this definition of participant wouldn't necessarily fly.
That is certainly true, and why the person element's description was
expanded, and why listPerson is available elsewhere.
> However, it seems to me
> unusual that if I should wish to introduce a <person> in the middle of
> the document, I would be obliged to wrap him or her in a <listPerson>
> element, and same with the <place>.
I hope you aren't confusing the names of people or places with
structured metadata about those people or places. <person> and <place>
are not meant for instances where those places are mentioned in the
text, but information you might point to from a <placeName> element. It
seems strange to me that you might want to dot individual person or
place metadata records casually throughout your document rather than
collecting them together. Gathering them together in one place in that
(or indeed another document entirely) strikes me as a much easier way to
keep track of them. However, I suppose I can envision sorts of compound
documents where you might want to do that, but grouping these with a
listPerson doesn't really seem that strenuous.
> I hope this doesn't come across as nit-picking. I just really want to
> get this right because while my students are working on the markup,
> I'm working on XSLT to produce Heml/RDF representation of the events,
> persons, etc., in order to blend this material with related data in
> CIDOC-CRM. So I want the XSLT to cover all eventualities, and I want
> the documents to be models of clarity.
I'm sure that others will have varying opinions from mine, but I'm
certainly happy to help to the best of my ability. When they have
produced these models of clarity, do provide links to them as sample
documents on the TEI wiki, same for the XSLT as well!
Dr James Cummings, Research Technologies Service, University of Oxford
James dot Cummings at oucs dot ox dot ac dot uk