From: Erik van der Poel (erik@vanderpoel.org)
Date: Mon Mar 07 2005 - 15:13:20 CST
Mark Davis wrote:
> If that link is clicked on, some user agent will eventually
> have to show http://badplace.com, and at that point it should be normalized
> in appearance. So double-struck C is not really a problem -- and it wouldn't
> be anyway, unless it looked like something *other* than a C.
Hi Mark,
I realize that Stringprep and Nameprep adopted the whole Unicode 3.2
character set partly because the General Category Values in the Unicode
Character Database did not appear to be stable yet at the time, and that
Normalization Form KC was adopted as is, partly because applications may
already implement that and it would take up more memory to have a
separate table for Nameprep, but now that we are discussing limiting
IDNs to Letters, Digits and Hyphen (LDH), which would require a separate
table anyway, it seems appropriate to question the original inclusion of
unnecessary characters like double-struck C, and to discuss ways to
exclude them. No?
Erik
This archive was generated by hypermail 2.1.5 : Mon Mar 07 2005 - 15:15:53 CST