Re: call for comments: Character Model Part II: Normalization

Phillips, Addison wrote:

> Two other parts, one on Normalization [2] and one on 
> Resource Identifiers [3] remain at lower levels of maturity.

[3] could be replaced by RFC 3987 everywhere.  Hopefully the
XML folks will at some point in time use proper STD 66 URIs,
and use the procedures specified in RFC 3987 to transform
IRIs into equivalent URIs.  Without RFC 3987 options or any
"legacy enhanced" cruft, just the clean 3987 algorithm, and
an understanding that percent-encoding in host names is not
for this universe, even if STD 66 isn't limited to DNS.

In [2] updating the references will be fun.  Please add the
new RFC 5198 while you are at it.  I consider RFC 5198 as a
kind of RFC 2277 amendment, roughly "we" (= IETF) "are not
only using UTF-8 everywhere a.s.a.p." (= until 2048) "as
stated in RFC 2277, we are actually using NFC" (as stated
in RFC 5198).

As you likely know I did not manage to get a normative [1]
reference in RFC 5137, it is only informative.  But for W3C
standards RFC 5137 should be pretty irrelevant.  AFAIK the
HTML5 folks intend to overrule [1] for their purposes of a
"Web browser implementor how-to", also not good for [1] :-(

 Frank

Received on Wednesday, 18 June 2008 22:12:30 UTC