Katakana Middle Dot again (Was: tables-06b.txt: A.5, A.6, A.9)

[John C Klensin]

--On Sunday, July 26, 2009 17:04 +1000 Wil Tan
<wil at cloudregistry.net> wrote:

> I'm fine with this, though I'd prefer "dirt simple" to be a
> plain "True". The advantage over (1) is that it allows room
> for explanation and warning to registries and developers, and
> over your proposed algorithm is that it doesn't prohibit
> labels that otherwise contain all Latin characters (decorated
> or not.)

Wil,

I have been silent on this for the last few days because I'm not
expert enough on Japanese (not expert at all) to evaluate
whether a given ruleset provides enough flexibility.   However,

* I thought we had agreed on Yoneya-san's proposal in April and
wonder if the additional discussion on this topic is a good use
of time.

* When making this PVALID is suggested, the conversation needs
to shift, at least in part, from the needs of Japanese to the
issues associated with having the character appear in
non-Japanese (or non-CJK) contexts.   In that context, while I
claim zero expertise in Japanese, I believe I'm adequately
competent to have opinions about a few "European" scripts and,
due to other work, to have competent opinions about visual
perceptions.  On that basis, and for the reasons below, while I
don't have very strong opinions about the details of the
contextual rule, I'm strongly opposed to making this character
PVALID.

Generalizing a bit leads me back to Harald's comment and perhaps
a guideline for thinking about these things.   Katakana Middle
Dot is a "Po" punctuation character.   We have banned _all_ of
those, modulo particular necessary exceptions -- a subject I
hope we don't have to reopen.  The issue here isn't using
CONTEXTO to add restrictions to a letter that would normally be
PVALID but using it to permit a character that would otherwise
be DISALLOWED to be used in limited contexts. 

I suggest that a useful meta-rule for thinking about this and
other ordinarily DISALLOWED punctuation characters is that the
only basis for allowing them is an argument that they are
sufficiently required in a particular context to justify an
exception and that they should therefore be CONTEXTx with the
rule reflecting that particular context and that context only,
not general discussions about where the character would be "safe
enough".  If nothing else, discussions about where some
punctuation character might be "safe" takes us back into a
discussion about whether we should DISALLOW punctuation
generally or whether we need to start a character-by-character
analysis for punctuation characters.  I'm pretty sure we don't
want to go there.

     john