RE: Protocol: Rolling in recent updates

[Hallvard B Furuseth <h.b.furuseth@usit.uio.no>]

Ramsay, Ron writes:
> I wouldn't have a SHOULD against octetStringOrderingMatch -
> semantically I can't see what sense this makes.

For octet strings that do not contain further structure, I imagine it
will often give the desired ordering - or at least approximately the
desired ordering.  caseIgnoreOrderingMatch doesn't do any better than
that anyway:

> Surely only characters strings have the notion of order.

They don't.  We assign them an ordering depending on content and usage.
E.g.:
Should '20' sort before or after '5'?
Should 'ä' and 'ö' sort after 'z' as in Swedish, or with 'a' and 'o'
as in German?
Should punctuation sort with space, or by the character set's
arbitrarily assigned character numbers, or by something else?

-- 
Hallvard