Re: Duplication: Remove most of section 3.3.2 and all of 3.3.4 ofdraft-ietf-ipsec-ikev2

[Paul Hoffman / VPNC <paul.hoffman@vpnc.org>]

At 11:43 PM -0400 6/17/03, Theodore Ts'o wrote:
>On Tue, Jun 10, 2003 at 08:05:43PM -0700, Paul Hoffman / VPNC wrote:
>>  Hi again. Most of section 3.3.2 in draft-ietf-ipsec-ikev2 (almost)
>>  duplicates the text from Jeff's algorithms document. Everything in
>>  the section after the Transform Type Values table should be removed.
>>
>>  All of section 3.3.4 of draft-ietf-ipsec-ikev2 is now covered in
>>  Jeff's algorithms document. Thus, the whole section should be removed.
>
>None of the code points in the IKEV2 document are authoratative; that
>honor belongs to the IANA registry.

Very true.

>   Yet, they are useful because
>there are a useful guide to initial implementors.

Correct.

>   It is for that
>reason that I believe a little duplication can be a good thing.

If you have to read two RFCs in order to implement IKEv2, having some 
duplication between them doesn't help anyone. And having differences 
between the two RFCs for the same values is exactly wrong.

What you are proposing is exactly one of the mistakes that we had in 
IKEv1 that we said many times we wanted to avoid in IKEv2.

>Otherwise, one could argue that all of the tables in IKEv2 are
>duplicative of the IANA registry, and therefore aren't needed; we
>should just establish the IANA registry and just delete all of the
>tables from the I-D (and from every other RFC that has initial
>codepoint values in them).

This isn't the way that the IETF works. New registries get 
established by RFCs that have values in them, and then can be 
expanded based on the rules established in the RFCs.

--Paul Hoffman, Director
--VPN Consortium