Re: [idn] rename ToUnicode?

[John C Klensin <klensin@jck.com>]

--On Monday, 13 May, 2002 10:55 +0000 "Adam M. Costello"
<idn.amc+0@nicemice.net.RemoveThisWord> wrote:

> Dan Oscarsson <Dan.Oscarsson@trab.se> wrote:
> 
>> IETF should refer to official standards organisations if
>> possible.
> 
> Correct me if I'm wrong, but I think ISO-10646 does not define
> normalization forms or case folding.  Therefore IDNA cannot
> avoid referencing Unicode.  Once we reference Unicode, there's
> no point in referencing ISO-10646, because it doesn't contain
> anything of interest that isn't already in Unicode.
> 
> Wouldn't it be ironic for the IETF to dismiss another
> standard-producing organization as not being official enough?
> Hasn't that same charge historically been made against the
> IETF itself?

Adam, for whatever it is worth, a distinction IETF has
traditionally made is between public-consensus, open-standards
bodies and consortia.  We've tried to be especially careful
about consortia that have high barriers to entry by interested
parties who are not large and well-funded corporations, since
such bodies are notoriously subject to capture by their member
companies.  We've also looked, at least implicitly, at the
stability of the materials we are contemplating referencing.
E.g., whether the organization, in practice, is committed enough
to stability to not change important elements of a referenced
document out from under us.

Note that the above paragraph doesn't say a thing about the
Unicode Consortium or UTC.  These are tradeoffs about which the
IESG and RFC Editor ultimately need to decide on a case by case
basis.   The same type of observation applies, I think, to your
position that, if we reference one UTC document, we might as
well reference Unicode documents exclusively and forget about
ISO 10646: Dan's preference, and probably mine (see below),
differ from yours, but it is important to see this as a
tradeoff, not a hard, fast, or obvious decision.

     john

p.s. If I were making the decisions around here, I'd reference
the code point set as "UCS" and then indicate, in the reference
text, a pointer to 10646 as authoritative and that the Unicode
spec contains the same set of code point mappings.  That is more
or less what the Unicode spec, as I read it, seems to say.  And
I'd say that as neutrally as possible.  But that is just my
personal taste in the matter.