On Thu, 18 Nov 2010 21:19:55 -0500, Ray Denenberg wrote
> I just want to be sure I'm clear: When you sort months and seasons 
> together, you are ok with a scheme where ALL seasons sort AFTER all months?

No. That's absolutely intolerable! We're sorting dates and not numbers! That's
would be equivalent to sorting 2000  before 1923-23-10. 

Its all about precision:
2000 is not before or after 2000-10
2000-10 is not before or after 2000-10-12
2000-24 would, similarly, not be before or after 2000-10-12.
2000-24 would also not be before or after 2000-10-11
(and so one, mapping what we assume about the start/end of the seasons without
the optional type modifiers)

2000-24 > 2000-23
2000-23 > 2000-22
2000-21 > 2000-1
2000-21 > 2000-2
(and so on, mapping what we assume about the start/end of the seasons without
the optional type modifiers)

I've suggested type modifiers such as those to specify meteorological etc.

Same with quarters that on additional thought I'd tend to not want to fold
into seasons so 25, 26, 27, 28 and with optional type modifiers)

2000-25 would be the first quarter of 2000. Without a modifier I'd say we
assume the start is Jan.

2000-25 is not before or after 2000-1
2000-25 is not before or after 2000-2
2000-25 is not before or after 2000-3


Note: Modifiers would not be secondary sorts but apply to the sort itself.

Should 25-1 denote the 1st normative fiscal quarter (defined as starting 1
April) then clearly

2000-25-1 > 2000-25
2000-25-1 is not before or after 2000-26

with more information we might have a fiscal year that starts in Jan. so we'd
have something like

2000-25-1-1 (2000-25-1-1 is, of course, equivalent to 2000-25 in sort)

2000-25-1-1 < 2000-25-1
2000-25-1-1 is not before or after 2000-25

Using this same system:
2000-25-1-2 would overlap 2000-25 and 2000-26. Typically one would sort it
then after 2000-25 but before 2000-26. etc. etc. etc.

Same as in sorting a Summer that starts 21 June. So the sake of this lets
assume that this is the default (just to save writing).

Its starting in the last days of the 3rd quarter so 

2000-22 > 2000-6
2000-22 is not before or after 2000-7

The only issue, as I mentioned in a previous post, is the season overlapping
the year mark due to the lack of agreed upon standardization for the
denotation of its year. In a library environment this should never pose a
problem since there is the additional information: a French journal arriving
in the Jan 2011 with the cover Winter 2011 would be encoded as 2010.

Loads of little details etc. need to be worked out...

The whole point of this exercise is to outline that we can define a sort that
works--- one that's perhaps easier to "standardize", I'd suggest, than sorts
of titles of books.

> --Ray
> -----Original Message-----
> From: Discussion of the Developing Date/Time Standards 
> [mailto:[log in to unmask]] On Behalf Of Bruce D'Arcus Sent: 
> Thursday, November 18, 2010 6:53 PM To: [log in to unmask] 
> Subject: Re: [DATETIME] seasons
> On Thu, Nov 18, 2010 at 6:05 PM, Ray Denenberg
> <[log in to unmask]> wrote:
> >> [mailto:[log in to unmask]] On Behalf Of Bruce D'Arcus
> >> So I need a way to describe the cover date for a periodical issue in
> >> ways facilitate two things:
> >>
> >> 1. sorting a bibliography reliably by date (vis-avis other entries that
> >> may have a month date)
> >
> > Bruce, I suppose I'm still confused about the requirement.  I take it that
you have periodicals of two distinct periodicities: monthly and seasonal. 
Sorting by month isn't a problem since there already is a sortable month
format, e.g. '2000-12'.  And if we allowed '2000-21' as a seasonal expression
then you could sort on that.  However it seems (from above) like you want to
sort bibliographies where some members are monthly and others are seasonal. 
That's not going to sort with this scheme.
> In my use case, there can definitely be cases where we need to sort
> items with month dates, and items with seasonal dates, in the same
> list. I would be fine with just defaulting to this scheme, so that
> Spring would sort after December in the list. But it would leave room
> for other sorting schemes as well.
> Bruce


Edward C. Zimmermann, NONMONOTONIC LAB
Basis Systeme netzwerk, Munich Ges. des buergerl. Rechts
Office Leo (R&D):
  Leopoldstrasse 53-55, D-80802 Munich,
  Federal Republic of Germany
Umsatz-St-ID: DE130492967