[dnsext] 2929bis DNS assumptions [was RRTYPE request: template for ZS record]

[Olaf Kolkman <[e-mail]>]

On Nov 21, 2008, at 11:02 PM, Jim Reid wrote:

>>
>> That suggests that the applications that use this RR need to be  
>> aware of the administrative boundaries in the domain namespace.
>
> They don't. They lookup a name (singular) and if they get a ZS  
> record back, that's it. If there's nothing there, they give up. No  
> searching or tree walking is involved whatsoever.

The name of the record and the specification made me wonder: question  
needed to be asked. Your answer is clear. It is probably good to add  
this clarification to the request before it disappears in the archives.

>
>
>> We do not have a good mechanism to find those zone boundaries and  
>> that indicates that there is more to this than simply query- 
>> response; there will be a bit of searching.
>
> There is nothing more than a simple query-response going on. It's  
> wrong to suggest or imply otherwise.

Although my mail did not contain question marks it was intended as a  
question ('if the above assumption is correct').

More to draft-ietf-dnsext-2929bis than to your request. I think that  
if a specification assumes knowledge of administrative boundaries in  
the domain namespace than that is an argument to deny the code point.  
I would think that such would be an incorrect assumption about DNS  
protocol behavior. Which could be, absent further specification, a  
reason to deny the code point. Is that a reasonable assumption?

--Olaf