M-label definition

John C Klensin klensin at jck.com
Wed Apr 8 19:37:38 CEST 2009



--On Tuesday, April 07, 2009 17:31 -0700 Paul Hoffman
<phoffman at imc.org> wrote:

> At 4:59 PM -0700 4/7/09, Lisa Dusseault wrote:
>>        "An M-label is a string that can be mapped into a
>>        [valid] U-label.  It may be a U-label, since those
>>        trivially map into themselves.  The category of U-label
>>        is a proper subset of the category of M-label." (Mark
>>        and others).
> 
> When I supported it, I was thinking of it as "...a valid
> U-label", that is, without the "valid" being optional.

FWIW, I put the "valid" in brackets only to emphasize it.  We
went over and over the various forms of "maybe it is a U-label"
many months ago and:

-- A U-label is, by definition, valid.  Anything that might be a
U-label, looks like a U-label but has not been verified to be
one, etc., is a "something else", but it isn't a U-label.  

	It is possible to think about a whole hierarchy of names
	of things between "some Unicode string" and/or "some
	Unicode string containing at least one ASCII character",
	at one end and "U-label" at the other.  If one were
	being precise, one could break Protocol down into steps,
	view most of them as a checking procedure that filtered
	out things that didn't pass the checks, and assign a
	term to each stage of getting closer to a real, actual,
	U-label.  

	I've been resisting what has felt to me like moves in
	the direction of establishing and defining those
	intermediate terms just because, while additional
	definitions add precision, they can also add confusion
	if their number reaches the point that people have
	trouble keeping track of them.

-- A "U-label" is convertible to and from an "A-label".
A-labels are, by definition, also valid, with the same set of
issues associated with "things that look like A-labels".

	The intention of the chart in Definitions was to try to
	get all of that nailed down.  Obviously, it was not 
   completely successful.

Does that help?
   john







More information about the Idna-update mailing list