Deprecated characters?

[John C Klensin]

--On Thursday, 17 July, 2008 08:05 -0700 Mark Davis
<mark.davis at icu-project.org> wrote:

> No it doesn't cause a problem: deprecation does not remove
> characters -- it just discourages their use.

But, Mark, if there are good and sufficient reasons to
depreciate a character, doesn't that implies that we should be
looking for ways to depreciate its use, especially for new
registrations?   Of course, a significant fraction of these are
irrelevant to IDNA2008 because of the prohibition on
compatibility characters and the requirement that strings be in
NFC form.   But it does seem to me that there is a more general
question here at least for the future and especially if the
reason for deprecating a character is to replace it with a
different preferred (or normalized) form.

Or am I missing something?

The question raised by the other proposal is whether our
requirements on strings that can be input to IDNA are strong
enough to prohibit ill-formed sequences entirely.  I believe
that the answer is "yes", but would appreciate it if someone
else would verify that.

best,
   john