WGLC: draft-faltstrom-5892bis-04.txt

[Andrew Sullivan]

[cc:s trimmed]

On Wed, Apr 27, 2011 at 04:49:18PM +0200, Stéphane Lancel wrote:

> I am sorry because I am not really expert in PVALID issues (changes, etc.),
> but I understood that PVALID once, PVALID for ever. I realize with this
> Draft that this means at character level not at codepoint level. Or am I
> wrong?

I think you're wrong in two ways.

First, it's true that the idea in IDNA2008 was that once something
would be PVALID, it would be PVALID forever.  But IDNA2008 depends on
Unicode and its properties.  If Unicode changes properties such that
something moves from PVALID to something else, then the desired state
is not achieved.

Now, one could just make a rule, "once PVALID then always PVALID."
But this is tantamount to sticking IDNA2008 to the Unicode version
that was around at the time of IDNA2008 publication.  We wanted to be
Unicode-version independent, and so that strategy doesn't meet the
goal.  

Second, I don't think "at character level not at codepoint level" is
right.  We're dealing in code points.

A

-- 
Andrew Sullivan
ajs at shinkuro.com
Shinkuro, Inc.