[rfc-i] Fwd: Re: Informational RFC to be: <draft-irtf-asrg-bcp-blacklists-10.txt>
rcallon at juniper.net
Tue Sep 27 11:41:24 PDT 2011
I agree with others that "experimental standard" is a contradiction. "Experimental protocol specification" is fine.
When I was on the IESG, there were a couple of discussions regarding whether non-standards track documents could use the RFC2119 keywords (SHOULD, MUST, and so on). Each time there was clear or even unanimous agreement that this is fine, and that the words are interpreted in the context. Thus if an experimental protocol says "MUST do <something>", then because the protocol is experimental then you don't have to actually do anything related to that protocol and you don't have to implement the protocol at all; However, if you implement the (experimental) protocol, then whatever that (experimental) spec has as a MUST is something that you MUST do if you want to claim that you have an implementation of that particular protocol.
From: rfc-interest-bounces at rfc-editor.org [mailto:rfc-interest-bounces at rfc-editor.org] On Behalf Of Bob Hinden
Sent: Tuesday, September 27, 2011 2:31 PM
To: RFC Interest
Cc: Bob Hinden; Dave Crocker
Subject: Re: [rfc-i] Fwd: Re: Informational RFC to be: <draft-irtf-asrg-bcp-blacklists-10.txt>
Excuse the top posting, but thought it easier to state my opinion.
I am fine with non-IETF standards track protocol specifications using RFC2119 key words because they have been come to be a good practice to develop interoperable implementations. That a good property for all protocol specifications.
Non-IETF standards track protocol specifications should be make it very clear they are not an IETF (or any other standards organization) Standard or an IETF Best Current Practice. IMHO, this document (i.e., draft-irtf-asrg-bcp-blacklists-10.txt) is not as clear as it should be in this regard.
On Sep 27, 2011, at 10:05 AM, Joe Touch wrote:
> On 9/27/2011 9:44 AM, Dave CROCKER wrote:
>> On 9/27/2011 9:38 AM, Joe Touch wrote:
>>> On 9/27/2011 9:30 AM, Dave CROCKER wrote:
>>>> As such, I think that non-IETF streams MUST NOT:
>>>> 2. Claim to conform to RFC 2119 (Key words for use in RFCs to Indicate
>>>> Requirement Levels)
>>> I disagree; tTspecifications too.
>> I don't understand what you mean. Please explain. You assert
>> disagreement but appear to be adding to my list rather than disagreeing
>> with it.
> It got cut off; it should have read:
> experimental protocols have specifications too.
>> Also note that "specification" is not really a formal document label in
>> our community.
> Yes, but experimental protocols have to be be able to say "MUST" or "SHOULD", etc.
>>>> 3. Have a title that asserts that the document is a standard or BCP
>>> I think "experimental standard" is OK in a title too.
>> Since we aren't even close to 1 April, I'll assume you are not kidding.
>> How could it possibly be acceptable to have a non-IETF stream assert
>> "experimental standard"?
> Individual submissions fall into this category.
> By "standard" I mean "specification" as per above - we don't have a good term for that, granted.
> The point is that "experimental protocol description documents" have a right - if not a responsibility - to refer to RFC 2119.
>> And by the way, what /is/ an experimental standard. If it's
>> experimental, it's not a standard, in our community.
> Yeah, yeah. An experimental protocol specification document.
> FWIW, I think of a standard as a document that defines a protocol, and an Internet Standard as the IETF track that's distinct - don't forget that not all standards come out of the IETF ;-)
> If we want to use a different word (specification, description, etc.), that's fine. The point overall is that experimental protocols NEED to cite RFC 2119.
> rfc-interest mailing list
> rfc-interest at rfc-editor.org
rfc-interest mailing list
rfc-interest at rfc-editor.org
More information about the rfc-interest