Re: Case sensitivity on the LHS

["Mark Davis" <mark.davis@xxxxxxxxx>]

>     I'm not sure what kind of mapping is done by NFKC... but I suspect the
> kind of problems that may ocur.
>
>     For example, if accents are removed from Latin letters once entered
...

NFKC does not remove the accents from Latin letters. If you are going to
comment on NFKC, pro or contra, you should comment on what it does, rather
than on simple speculation as to what it does. The formal results are in UAX
#15 on the Unicode site. There is a chart that shows the effects on
characters on http://www.unicode.org/charts/normalization/. For example, if
you look at Latin characters on
http://www.unicode.org/charts/normalization/chart_Latin.html you will see
that "ô" remains as "ô" in NFKC.

Mark
________
mark.davis@xxxxxxxxx
IBM, MS 50-2/B11, 5600 Cottle Rd, SJ CA 95193
(408) 256-3148
fax: (408) 256-0799

----- Original Message -----
From: "Alain LaBonté" <alb@xxxxxxxxxxxxxxx>
To: "Simon Josefsson" <jas@xxxxxxxxxxx>; "IETF IMAA list"
<ietf-imaa@xxxxxxx>
Cc: <alb@xxxxxxxxxxx>
Sent: Tuesday, February 11, 2003 05:58
Subject: Re: Case sensitivity on the LHS


>
> A 14:17 2003-02-11 +0100, Simon Josefsson a écrit :
>
> [non-quoted correspondent]
> > > You claim that "many distinct names" of humans will get collapsed by
> > > NFKC into the same name.  So far you have provided zero examples.
Could
> > > you supply a few more please?
> >
> >[Simon] If you want to modify existing standards, I think you need to
prove
> >that it doesn't break things, not the other way around.
>
> [Alain]  Could somebody summarize what is the actual behaviour of NFKC for
me?
>
>     I'm not sure what kind of mapping is done by NFKC... but I suspect the
> kind of problems that may ocur.
>
>     For example, if accents are removed from Latin letters once entered
> (which in most cases would be convenient for those who can't enter
accented
> characters -- because I would like them to reach me if my email address
> were to be « Alain.LaBonté@iquébec.com »), there might indeed be
collapses,
> and I will give an actual example, using a look-alike family-name cluster
> from the city of Québec's telephone book:
>
>     There are actual real-life collisions for example, in these cases:
>
>     Cote B
>     Côte B
>     Coté B
>     Côté B
>
>     This was already one of my favorites for dictionary ordering (the 4
> family names, are 4 disting French words, meaning respectively "quote",
> "hill side", "quoted" and "side").
>
>     However this case of collapse is imho no different from cases of names
> collapsing just because they are simply identical. One has, with the same
> ISP, to find extra ways to distinguish them. Imho an email address
> "B.Cote@..." should be able to reach "B.Côté@..."
>
> Alain LaBonté
> Québec
>
>