EDTF assumed neither-- at least from my perspective (the archive is filled with the discussion)--- an astronomical
definition of season nor a "northern hemisphere" localization. We discussed this at quite a bit depth.
The problem is:
- when a source says "Spring 2018" we don't know quite what semantics have been attached.
The first day of Spring, for example, was traditionally viewed as 1 Feb in Ireland.
(these and a multitude of other examples I discussed here years ago-- c.f. archive)
See also https://en.wikipedia.org/wiki/Spring_(season)
- As has been pointed out we can't quite be sure of the span of an expression of "Winter 2017" in the Northern
Hemisphere. Does Winter span from 2016 to 2017? When is "Autumn/Winter 2017"?
Throw in hemisphere and things get even wilder.
While we could define rules and semantics for Winter, Spring, Summer and Fall we would often need
significantly more information (and probably a higher precision of expression) than might be available
to map from one expression of "Spring 2017" into a canonical "Spring" (or season) and year.
(the suggestion that 21 is the season that one occupies in the 1st month of the year, etc.)
- We don't know where the expression has been uttered. This is extra information.
- Sometimes it is even quite ambiguous (ill-defined). "Larry called home to tell his mother in New York how much
he was enjoying the Summer in New Zealand."
(Summer from the perspective of the New Yorker when at home in New York or from New Zealand".
The expression of a season can't (and should never) be mapped to some other date specification that
does not use seasons-- unless, of course, one could resolve the date into say day precision).
Since we did not want to get too involved in this we choose the best way out: leave the seasons as ambiguous.
(P.S. this is also where the problem with ranges defined in mixed precision between seasons and months lies.)
The expression "Winter 2007" just gets mapped to 2007/24.
If I recall we did discuss adding some additional extensions to localize things-- similar to timezone (unfortunately
also significantly more complicated to express and encode).
(and remember that the span of timezones in the world is > 24 hours so the same instant can be reported, lacking the
Zone information, not just on different days, resp. month, but even year. Edge cases matter!)
> -----Original Message-----
> From: Discussion of the Developing Date/Time Standards
> [mailto:[log in to unmask]] On Behalf Of Tony Benedetti
> Sent: Thursday, January 18, 2018 9:56 AM
> To: [log in to unmask]
> Subject: Re: Seasons in intervals (ISO 8601)
> >>> Christoph Päper
> >>> EDTF may have been developed for bibliographic purposes where the
> astronomical, Northern-hemispheric definition predominates, but ISO
> 8601 has a much broader scope.
> You're absolutely correct Christoph -- EDTF obviously shares that
> "broader [non-bibliographic] scope" with ISO 8601. Otherwise, years
> would be limited to 4 digits ... that is, unless we think people will
> be cataloging periodicals that will be published in Y17E7 (170000000)
> ... or books that were published way back in Y-17E7 (-170000000). Both
> of the preceding example dates were taken from "4.5.2 Level 2" of ISO
> 8601-2. Those sorts of extremely distant dates are clearly meant for
> historians, curators & scientists and not to accompany ISBNs & ISSNs.