Re: An argument against multiple character sets

["Martin J. Duerst" <duerst@w3.org>]

Let's take the Swedish A with ring above. Unfortunately,
there is more than one way to represent that. One is the
usual way, U+00C5. The next is ANGSTROM SIGN, U+212B.
And then is a plain A followed by COMBINING RING ABOVE,
U+0041 U+030A. The Unicode standard defines that these
are all exactly equivalent (canonically equivalent).
The user won't be able to distinguish them.

Assume the worst case: all three are registered, each
client does something different (i.e. sends either of
these as it pleases when the user inputs an A with a
ring above). Such a DNS is clearly not usable.

Regards,   Martin.

At 13:14 00/01/26 -0500, J. William Semich wrote:
> If all domain names are registered using a single universal standard
> encoding then all registered domain names will be unique within a
> particular TLD, assuming the current requirement for a single root server
> system continues to exist.
> 
> What's the problem? Can someone provide some examples? Sorry if it's not so
> obvious to me. <smile>
> 
> Bill Semich
> .NU Domain
> 
> At 06:33 PM 1/26/00 +0100, Dirk-Willem van Gulik wrote:
> >On Wed, 26 Jan 2000, Martin J. Duerst wrote:
> >
> >> > This because of the impact on registration
> >> > issues, because domain names have to be "unique" in a zone at time of
> >> > registration. Because of that, the rules for equality have to be known at
> >> > the time of registration,
> >> 
> >> Not only the rules for equality, also the ways to deal with this
> >> equality. If some equality is dealt with on the client side, then
> >> registration has to make sure the right variant of the name is
> >> registered.
> >
> >But if all equality can be dealth with on the server (even
> >if this means the server gives back a list in some cases;
> >and leaves that to the client), then you only need to define
> >the comparison function within the spec; and you can leave
> >the remainder to the client/UI/implementation.
> >
> >Dw
> >
> >
> 
> 


#-#-#  Martin J. Du"rst, World Wide Web Consortium
#-#-#  mailto:duerst@w3.org   http://www.w3.org