On Wed, 24 Nov 2010 15:13:46 +0000, Simon Grant wrote
"If we had a way of signifiying "one value from within", then putting that
together with the range notation (whatever we agree it finally to be,"

But they are not really ranges but precisions, fuzziness, uncertainty and
not-yet collected...

Ranges are quite well defined in the ISO specification. Technically also year,
month, week, day and hour precisions are specifiable. But we want, perhaps
other precisions such as decade and century.. and some fuzziness and
uncertainty. That's where these ?, u, ~, [] and {} enter the scene for date.

These can then be applied to ranges as well....

> I do think we need an alternative to ISO for a range. In particular, I think
we need external delimiters -- some sort of brackets -- enclosing the range.
This could go with the ... syntax as an alternative to ISO "/". I don't
currently have any positive suggestions for how to represent "one value from
this range" except that it should use exactly the same syntax as the range. 

We already have Glob with patterns such as 

[123] to match 1, 2 or 3
[0-9] to match 0,1,2,3,4,5,6,7,8 or 9
[^123] to match anything but 1,2 or 3.

and () for grouping.

From 314:

[1667,1668, 1670-1672]
One of the years 1667, 1668, 1670, 1671, 1672 

From 315:
{1667,1668, 1670-1672}
All of the years 1667, 1668, 1670, 1671, 1672 

In 313 we have:

December 25 sometime during the 1560s ('?' := questionable)

to contrast to:


Which would be December 25 in the 1560s, but it is not certain which year ('u'
:= unknown)

We also have the character  '~' for "approximate".

In 305:


In ISO we have weeks (in the year) so this can apply not just to day, month,
year, hour, minute, second but also week.

The question is: Do we want to allow (as in Glob) things like:

2005-06-1[5-8]  (as equivalent to 2005-06-[15-18] )



And how about


[Feb has only 28 or, in leap years, 29 days so there is no 30 Feb and no month
has 32 or more days]


 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