[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [idn] Comparisons of the proposals



I have an implementation question with respect to the comparisons process.
I'll admit I'm not that familiar with the implementation of the UTF-5 on
the server side or the CIDNUC solution in general, and it may be that this
question is totally off track. If this is a stupid question, please bear
with me....

But I am familiar with the UTF-8 solution, and know that it has one
important characteristic - ASCII transparency - with regard to backward
compatibility with current ASCII Top Level Domains and all other ASCII
domains and host names on the Internet. ASCII transparency means you don't
need to encode ASCII in order to use it with the UTF-8 encoding method on
the client or server level.

So using UTF-8, the root servers would continue to use the un-encoded ASCII
characters .com for all .com domains, .net for .net domains, and ditto for
any other ASCII Top Level Domains (as would all other name servers on the
Internet). The same holds for browsers and other applications that use
ASCII domain names.

And, if so enable, the client and the server would use UTF-8 encoded names,
like .domän for .domän (that's dom-a-doubledot-n), in those cases where an
idn is being used. These two "types" of encoding domains (ASCII and UTF-8)
could co-exist and resolve inside any UTF-8 enabled name server (like
NUBIND). And the current .com, .net, etc. domain names would continue to
resolve across the Internet using their ASCII representations, without
anyone having to change their name server software to support idn (if they
chose not to) and without having to enter UTF-5 or CIDNUC encoded versions
of all the current ASCII top level domain names, second level domain names
and host names on the Internet.

My question is, how do UTF-5 and CIDNUC deal with that?

In other words, if a root server were to support both .com and .domän using
UTF-5 encoding, for example, would both TLDs have to be encoded in that
root server's zone files using UTF-5, so .com would become .M3MFMD and
.domän would become .M4MFMDU4ME? And would each browser have to query those
name servers using the UTF-5 encoding for *all* ASCII .com domain names as
well?

Would every other name server on the Internet that uses .com have to change
its zone files and use the same UTF-5 encoding .M3MFMD for ".com"?? 

And would every domain and host name on top of .com have to be re-entered
in its complete UTF-5 encoding, such as
N7N7N7.MEM5N4N7MFN2MBN3MFMCN5N4M9MFMEN3.M3MFMD for www.networksolutions.com?

And would all current ASCII representations with meaning in our zone files
have to be replaced with the encoded versions??

In the same vein, how would it work for CIDNUC?

Thanks to all for any additional info or for clarifying any confusion on my
part...

Bill Semich
.NU Domain

Bill Semich
President and Founder
.NU Domain Ltd
http://whats.nu
bill@mail.nic.nu