[idn] "server-side" implementation

[Dave Crocker <dhc@dcrocker.net>]

Folks,

Since the idea of restricting iDNS to require only implementation on the 
server has been appearing a number of times.

We need to shoot this idea dead.  It has nothing to do with networking or 
the IETF.  If something is restricted to be within a machine, it is not 
networking.

Unfortunately, iDNS requires sending an "international" string between a 
DNS client and a DNS server.
That IS a networking activity, and it means that there must be a protocol 
between client and server, supporting that international string, such as UTF-8.

So anyone who thinks that iDNS can be restricted to a server needs to 
consider how the iDNS string shows up at the server.

d/


>I feel a server side implementation is more appropriate for the long run, 
>and to my understanding there are companies that already have technologies 
>that do similar transformations and will work for international user, by 
>simply setting up root servers that will resolve those names, and in worst 
>case senario, the client has to set the networking of their machine to 
>query a name server that can convert the translated string.
>
>So again I do not see why we are discussing Walid's patent??  Even if we 
>have a majority of people here

----------
Dave Crocker   <mailto:dcrocker@brandenburg.com>
Brandenburg InternetWorking   <http://www.brandenburg.com>
tel: +1.408.246.8253;   fax: +1.408.273.6464