Remark: Started today the road to change the -- to /. Dont know where
this tradition of -- comes from but unfortunatly I think it will take
some time before the change is in place. Good to have some reasons why
the use of -- isn't the way to go.

I've got work to do to change that tradition then....


>>> Ray Denenberg <[log in to unmask]> 2011-03-09 21:10 >>>
From: Karin Bredenberg
> For the double dash:
> Here in Sweden its a tradition when you write an archival
> and are making an interval to use the double dash and even the
> dash is sometimes used. But when you state the normalization you use
> the slash for period.
> Example one from the ead-homepage
> ( 
> <unitdate type="inclusive" normal="1952/1964">1952-1964</unitdate>
> Example two from the swedish description
> ( page 6):
> <unitdate certainty="[X]" type=”[X]”>[YYYY]--[YYYY]</unitdate>

I'm sorry but I am now convinced that we cannot adopt the double dash
form for an interval.

In general, and in the interest of moving more quickly to closure, I
would like to advance the principle that we do not adopt an alternative
form of expression simply because some other syntax uses it.

But in the case of double dash there is a much more compelling reason,
one that just occurred to me:

Double dash is used to mean something entirely different in ISO 8601. 

'--01-31' means "January 31" (year unspecified).
'--01' means "January. 

And while these features (currently 104 and 106) are on last call and
on the verge of elimination from our spec, just because we profile out
some IS0 8601 feature from, we can't simply adopt the syntax from that
feature to use with different meaning.

Bruce  - I am not sure what this implies for RIS, we need to think it
through.  The use of consecutive dashes as discussed for RIS seems
consistent with its use in 8601.