[tei-council] Comments on Place proposals from TEI
Lou Burnard
lou.burnard at computing-services.oxford.ac.uk
Sun Apr 22 11:59:34 EDT 2007
Dear Tom
Many thanks for your timely comments. I've alerted the Council to your
wiki page and also I think we have a link to it from the relevant trac
item for our own agenda.
Here are my personal reactions to your comments, not to pre-empt the
further discussion I expect to take place next week :
1. If accepted, the testplace proposals will probably be integrated
into the existing chapter on Names of Places and Persons which has also
undergone quite a lot of revision as you note. You might like therefore
to point to the current state of that chapter which is (as of this
morning) at
http://tei.oucs.ox.ac.uk/Guidelines/Source/Guidelines/en/guidelines-en.xml.ID=ND
rather than the unrevised last stable version -- this is dynamically
generated from the svn repository in which changes are being made at a
great rate of knots
2. Thank you for the nice remarks about what distinguishes the TEI from
the rest. You're right that we need to be clear about why the TEI goes
down this particular muddy well trodden track at all, and certainly to
alert readers of the Guidelines to other possibilities where they exist.
3. Named time periods. The treatment of dates and times has just
undergone major surgery in order to support various kinds of automatic
processing and validation, mappings to both ISO and W3C date formats
etc. When the patient is mobile again, I agree that we need to start
thinking about "named" dates. It should not be too difficult.
4. Linking assertions with responsibility statements is a recurrent
thread. We have discussed a generic <assert> element, to combine an
assertion with an indication of responsibility and certainty, but backed
away from it as being too complex, and not adding much beyond what the
existing @resp and @cert attributes provide; also there is an existing
method for marking certainty
(http://tei.oucs.ox.ac.uk/Guidelines/Source/Guidelines/en/guidelines-en.xml.ID=CE)
which should be taken into consideration.
5. Adding GeoRSS is definitely an option. We only specified GML because
that was the one we found on the web... no intention of being exclusive.
6. <relation> vs containment. It always makes me uneasy to have too many
ways of doing exactly the same thing, but it's hard to see what
containment could mean for <place>s within <place>s other than what is
proposed, nor to argue against allowing for relationships other than
containment, so we wound up with both. Contrast this with the discussion
of what <nym> within <nym> means -- something quite different.
7. <locale> was a pre-existing TEI element which we pressganged into
service here. It does seem a bit problematic as to name, and its overlap
with other ways of characterizing a place.
8. Ethnica sounds interesting. But I think it's a kind of placeName
still, not a new kind of name, even though its location may be hard to
pin down. After all whether we think of Bristol as "the place with a
bridge" or "the place where the Bristolians hang out", it's still a place.
9. Absolutely no intention or need for TEI to re-invent detailed
geographic metadata already done better by GML and similar. We could do
with some real examples though, as you note.
10. SpatialML is an interesting one which I have glanced at but not
studied in detail: it comes from the very different world of "named
entity recognition" . Certainly we ought to review it and the others
you mention, before rushing ahead, if only as a fruitful source of examples.
best wishes
Lou
More information about the tei-council
mailing list