[tei-council] processing TEI times

[Syd Bauman]

2005-09-22T11:27+01, Sebastian said:
> On dates/times again, one should bear in mind that XML processors
> may start using datatype information. At the moment, the only thing
> where it matters is in validation; but in the world of W3C Schema
> and XSLT2, the processing application can say "what basic datatype
> did you work out for this attribute". So if our schema said
> "xsd:dateTime | long-regexp-allowing:13:15as-a-time", then the
> processor might take different actions depending on whether you did
> 13:15 or 13:15:00.
> It is not easy to think through whether this matters or not.

2005-09-22T16:31+01, Sebastian said:
> you may one day wish to do a proper sort of TEI-encoded elements
> according to their date. If they emerge into the PSVI as having
> matched a regexp, you'll be at a disadvantage.


I've decided that I (partially) agree with Sebastian. I still believe
whole heartedly that it is incredibly important to be able to record
times that are less precise than to the second, and I still believe
that permitting the value= attribute to directly reflect this reduced
precision is the right way to do this. However, after poking around a
bit I've concluded that it *is* quite possible, if not likely, that
users will run up against software that handles times precise to the
second just fine, but barfs on less precise times.

Thus, I've written a stylesheet for converting a TEI P5 document with
<time> and <date> elements that have reduced precision times on value=
to a similar document with (falsely) precise times. Actually, I've
written two: one for XSLT 1.0 and one for XSLT 2.0 processors. These
stylesheets can be found on the TEI wiki:
  http://www.tei-c.org/wiki/index.php/InsertFalsePrecision1.xsl
  http://www.tei-c.org/wiki/index.php/InsertFalsePrecision2.xsl

Lastly, I've updated the pattern in tei.data.temporal to accept times
precise to only the hour, and imprecise times that also have dates.
Furthermore I've updated the pattern to allow only W3C-legal timezone
designators.

Because the pattern has gotten quite unwieldy, I've also written a
tiny test schema and test file for it. The test schema includes
annotations to explain the sub-patterns. They can be found at
   http://www.tei-c.org/Council/test_time_pattern.rng
   http://www.tei-c.org/Council/test_time_pattern.rnc
   http://www.tei-c.org/Council/test_time_pattern.xml
In order to make it easier for a Council member to read the
annotations, I've put them up separately at 
   http://bauman.zapto.org/~syd/ttpe.html