[Standards] Re: IDNA text for rfc3920bis

[Joe Hildebrand]

On Apr 16, 2007, at 12:35 PM, Mridul wrote:

>   I think the query was because :
> If you used only hostnames - then they get resolved to the default  
> domain : like peer1, peer2, etc would resolve to peer1.domain and  
> work.
> With peer1.localhost, unless you have resolution entries for these  
> in the dns, it would not work (and other than /etc/hosts style  
> entries - which are a pain to maintain across nodes).

According to RFC 2606:

<blockquote>
       The ".localhost" TLD has traditionally been statically defined in
       host DNS implementations as having an A record pointing to the
       loop back IP address and is reserved for such use.  Any other use
       would conflict with widely deployed code which assumes this use.
</blockquote>

So, "foo.localhost" pointing to another machine seems right out.

> For argument sake, we could envision that for high availability  
> nodes ... use fully resolved hostnames.

I have use cases where this won't work.  I can't share them at the  
moment, unfortunately.  Suffice it to say this restriction would make  
large numbers of deployed systems effectively non-standard in one  
fell swoop.  As such, I would strongly object to this change in the  
bis drafts.  I *would* support a best-practices XEP, however.