apologies for making yr head spin. i dont disagree with any of that summary. The problem is in the following example which appears in att.datable.w3c <egXML xmlns="http://www.tei-c.org/ns/Examples" source="#fr-ex-Laclos-liaisons"> <opener> <dateline><placeName>Paris,</placeName>, ce <date when="--12-11" notBefore="1699-12-31" notAfter="1799-12-31">11 décembre 17.. </date></dateline> <salute>à Madame De Rosemonde.,</salute> </opener> </egXML> I think this example is trying to be too clever. It provides two normalisations of the date "11 decembre 17..", one using @when, the other using the combination of @notBefore and @notAfter. Either on its own would be fine, but both together is needlessly confusing. I'd suggest simply removing the @when one, since we've already got examples of that way of indicating an imprecise date. On 16/03/16 21:33, Syd Bauman wrote:
Lou and Hugh, you've made my head spin and in large parts seem to be somewhat off the ranch. I believe you wholeheartedly if you say I made a mistake, but could you point out exactly what that mistake is?
To recap the current released status:
* @from & @to designate a duration
* @notBefore & @notAfter designate a range during which something happened (that thing may have happened instantly or over a duration)
* The Guidelines do not explain what any of the following combinations mean: - @from & @notBefore - @from & @notAfter - @to & @notBefore - @to & @notAfter - @when & any of @from, @to, @notBefore, @notAfter
To recap the issue:
* I complained that an example in the GLs used a combination for which there was no meaning; I thought we should change the example so it used a combination we explain.
* Council said "No, Syd, that's a perfectly valid combination, the prose of the GLs should be changed to explain what each combination means -- we need a table".
* I said "OK, I'll go do that", not remotely realizing how hard it is to generate a decent looking table.
To recap the discussion:
* James, at least, and I think one or two others, asserted (w/o objection) that any combination that includes @from or @to is definitionally a duration, not a range, even if @notBefore or @notAfter is present.
* Peter, James, and I each asserted (initially w/o objection) that any of @from, @to, @notBefore, or @notAfter makes no sense with @when; therefore those combinations should not be permitted.
* Later Hugh pointed out that there is an example with the use of @when, @notAfter, and @notBefore in the GLs -- so maybe those combinations should be allowed.
To recap my changes:
* Added the table to explain what each of the following 2-attribute combinations mean - @from & @to (duration) - @notBefore & @notAfter (range) - @from & @notAfter (duration, uncertain endpoint) - @to & @notBefore (duration, uncertain beginning)
* Added prose to explain what the following 3-attribute combinations mean - @from, @notBefore, @notAfter (duration, uncertain endpoint) - @to, @notBefore, @notAfter (duration, uncertain beginning)
* Added some examples
I hope to address Hugh's post on the ticket (#1331) in a few mins.