Re: [TOOLS-DEVELOPMENT] Draft of inline-errata SoW

[Eric Rescorla <[e-mail]>]

On Thu, Jan 31, 2019 at 3:18 PM Heather Flanagan <[e-mail]> wrote:

>
>
> On Jan 31, 2019, at 12:44 PM, Eric Rescorla <[e-mail]> wrote:
>
>
>
> On Thu, Jan 31, 2019 at 12:20 PM Heather Flanagan <[e-mail]>
> wrote:
>
>> Hi ekr,
>>
>> I understand your point, but must disagree - these documents will not be
>> fully archived,
>>
>
> I don't see how this is relevant. In my hypothetical, the version which is
> "fully archived" is simply wrong (and in fact, self-contradictory), so it's
> not sensible to talk about having normative status on that point.
>
>
> It is relevant because the RFC Series is as much about archiving formally
> approved documents as it is making those documents available.
>

I don't agree that it's as much about that -- though I agree it's partly
about it -- but I suppose we can leave this for another day.


The archived documents may include information which has proven to be
> wrong, but the information is not corrected until we have a new document
> published.
>

This doesn't seem entirely accurate. The information *is* corrected in that
we file errata and have them approved and available, and those errata have
wha What we're talking about now is exactly how we present those errata.


>
> and the verification process is outside the consensus process (which I
>> think is important for IETF documents).
>>
>
> This seems like an interesting point, but is actually mitigated by the
> fact that if the approver makes a mistake, we can change the erratum again.
>
>
>
> Yes, true, but that is isn’t required to flow through any kind of
> consensus process.
>

Well, this is true, but it's also the case that these kinds of changes are
regularly made in AUTH48 without going through any kind of consensus
process, with only the IESG and the authors serving as the arbiters of
whether a change needs consensus. The level of consensus isn't materially
different just because the RFC has been published.


Taking a step back: an RFC -- at least a protocol specification --  is a
> set of instructions for how to implement a protocol. And what people need
> to know is how to actually implement the protocol. And when there is a
> clear error in the specification that doesn't match how the specification
> is being interpreted in practice, then it isn't helpful to act as if the
> error has normative force. The question is what the best interpretation of
> the RFC is at a given time.
>
>
>
> I get that, and am very sympathetic to that point. That’s why I’m in favor
> of creating these annotated RFCs in the first place, as an aid to both
> implementers and future -bis authors. But the annotated RFCs are a moving
> target with an insufficient consensus process for change, making them
> something I’m not willing to include as an official, normative output
> within the RFC Series.
>

It seems to me that the questions of what has normative force is really a
question for the IETF, not the RFC Editor. The IETF (with consensus, of
course) could decide tomorrow to mint a new class of documents that had
normative force and send them through some entirely separate publication
process (Github, Web pages, engraved on the face of the moon) and they
would still have whatever normative force one ascribes to the IETF
consensus process. Of course, that's a separate question from what goes in
the RFC Series.

All in all, I think this suggests that it would be better to have this
material on the IETF site and let IETF make whatever statements it wants to
about the status of the errata.

-Ekr


>
>
>> -Heather
>>
>> On Jan 31, 2019, at 11:29 AM, Eric Rescorla <[e-mail]> wrote:
>>
>> Hmm... I'm not sure that that's quite the desirable outcome.
>>
>> Let's take the example of a clear (and minor) technical error: in the
>> body of the text, the RFC says that we use code point X and in the IANA
>> considerations it says code point Y and the code point that people use is
>> actually "Y". So, someone files an erratum that changes the body of the
>> text to read Y. In this case, we absolutely do want people to think that
>> the annotated version is the one with normative force.
>>
>> -Ekr
>>
>>
>> On Thu, Jan 31, 2019 at 9:28 AM Heather Flanagan <[e-mail]>
>> wrote:
>>
>>> Hi Sandy,
>>>
>>> You know I like the idea of a sandbox space - the RFC Editor needs a
>>> better way to explore ideas in a visible way. That said, there is concern
>>> about any hint of normative-ness to the annotated docs, and hosting them on
>>> the RFC Ed website might suggest some amount of officialness. I think we
>>> can potentially mitigate that by including something in the CSS that THESE
>>> ARE NOT NORMATIVE DOCUMENTS, but other ideas are welcome.
>>>
>>> -HEather
>>>
>>> On Jan 30, 2019, at 6:47 PM, Sandy Ginoza <[e-mail]> wrote:
>>>
>>> Hi all,
>>>
>>> I figured this comment/question would be more appropriate here rather
>>> than as feedback for the SoW.  I can’t remember if placement was decided
>>> when the IESG discussed this topic at the last IETF, so sorry if this is
>>> re-raising discussion I should already be aware of.
>>>
>>> The text says:
>>>
>>> This tool will initially be used to create pages served somewhere other
>>> than the RFC Editor’s website (www.rfc-editor.org). After gaining
>>> experience and community feedback, the pages might be moved to that
>>> website. Thus, the development of the tools should assume that the way the
>>> results are deployed will change over time.
>>>
>>> The RFC Editor has been talking about creating an experimental space on
>>> our site.  I think it’s worth considering having these annotated documents
>>> live in an experimental space on the RFC Editor site so they’re on the site
>>> from the beginning.  This seems like a perfect use for an experimental
>>> space.  Also, I believe that once these documents live somewhere, that is
>>> where users will look for them; if they become the de facto RFCs, I think
>>> we want them to be on our site.
>>>
>>> Thanks,
>>> Sandy
>>>
>>>
>>> On Jan 29, 2019, at 12:07 PM, Robert Sparks <[e-mail]>
>>> wrote:
>>>
>>> Adam's developed an initial draft for the inline-errata SoW. He started
>>> it as a google-doc and my initial set of feedback changes were easy enough
>>> to make there. Unless someone objects, I suggest we continue the
>>> development of this one using that tool. If you do object, for any reason,
>>> please let me know and I'll send a PDF while we adjust.
>>>
>>>
>>> https://docs.google.com/document/d/1T-6eBrNPC7aIxOTChdIZE7fecbxUJ7kq-F6RhyCNUGI/edit?usp=sharing
>>>
>>> RjS
>>>
>>> _______________________________________________
>>> TOOLS-DEVELOPMENT mailing list
>>> [e-mail]
>>> https://www.ietf.org/mailman/listinfo/tools-development
>>>
>>>
>>> _______________________________________________
>>> TOOLS-DEVELOPMENT mailing list
>>> [e-mail]
>>> https://www.ietf.org/mailman/listinfo/tools-development
>>>
>>>
>>> _______________________________________________
>>> TOOLS-DEVELOPMENT mailing list
>>> [e-mail]
>>> https://www.ietf.org/mailman/listinfo/tools-development
>>>
>>
>>
>