Re: [dnsext] alternatives to the proposed NEGATIVE pseudo-RR

["George Barwood" <[e-mail]>]

----- Original Message ----- 
From: "Paul Vixie" <[e-mail]>
To: "George Barwood" <[e-mail]>
Cc: "Jim Reid" <[e-mail]>; "Olaf Kolkman" <[e-mail]>; 
"Stuart Cheshire" <[e-mail]>; "Wouter Wijngaards" 
<[e-mail]>; "IETF DNSEXT WG" <[e-mail]>; 
<[e-mail]>
Sent: Thursday, September 25, 2008 1:11 PM
Subject: Re: [dnsext] alternatives to the proposed NEGATIVE pseudo-RR


>> Has defining a class been considered?
> this isn't what classes are for, and it isn't how classes work, so, no.

The standard doesn't state the purpose of classes, it just enumerates them.

mRDNS is already using class for another purpose.

>> This seems cleaner to me than using a Type.
>
> rdata format is class independent.  rdlen=0 would be a protocol violation.

A and AAAA are specific to IN, so may have other formats in other classes 
surely?

Still,  class seems dead for practical purposes, so I concede that reviving 
it is probably not a good idea.

> further, the NSEC format is actually much more compact for this purpose.
> (the responder in stuart's case doesn't know what rrtypes the initiator
> may be contemplating, so it's nec'y to enumerate the full set and 
> illuminate
> those present/absent.) 



--
to unsubscribe send a message to [e-mail] with
the word 'unsubscribe' in a single line as the message text body.
archive: <http://ops.ietf.org/lists/namedroppers/>