Re: Handling IPR disclosures through draft reorganizations

[Brian E Carpenter <[e-mail]>]

On 24-Jul-23 17:04, Joel Halpern wrote:
> I agree that the rules say that, and that was intentional on the part of
> the community.  So it should be applied by the tooling.
> 
> There is a potential implication.  I would hope that we would not expect
> corporate legal departments to track the changes in the drafts and file
> separate disclosures against each revision.  That would seem to create a
> lot of work for little benefit to the community.  I do not believe that
> was the intention when the rules were written.

That is logical, so the question is really whether the long-written
rule needs to be changed (but in a way that answers Carsten's
question: what happens when a draft bifurcates?). Also there's
probably a need for a retroactive rule for this period when
the tool and the rule have been discordant.

Perhaps we should complete the following two sentences:

"IPR disclosures that do not indicate a specific version
number are deemed to apply to ..."

"When a draft is updated, renamed, split, or transformed
into an RFC, ..."

    Brian


> 
> Yours,
> 
> Joel
> 
> On 7/23/2023 7:52 PM, Brian E Carpenter wrote:
>> Jay,
>>
>> Are you aware that the IPR disclosure page apparently does not enforce
>> the very precise rule in BCP 79 that I-Ds cited in disclosures must be
>> referenced by specific version number?
>>
>> This seems to be a 9-year-old bug, with consequences as described in
>> Carsten's original message at
>> https://mailarchive.ietf.org/arch/msg/ipr-wg/Vr13La4lbl5Ck61ISlPC20V1Hmo
>> .
>>
>> Regards
>>      Brian Carpenter
>>
>> On 20-Jun-23 09:08, Brian E Carpenter wrote:
>>> I believe that Carsten has uncovered a very serious situation.
>>>
>>> On 19-Jun-23 19:50, Carsten Bormann wrote:
>>>> In July 2021, the CoRE WG split the draft-ietf-core-dynlink into two
>>>> parts, draft-ietf-core-dynlink proper (about dynamic link bindings),
>>>> and the separately progressed draft-ietf-core-conditional-attributes.
>>>> It has been slow going, but we now have mostly completed work on the
>>>> latter, and would like to take up finishing the former.
>>>>
>>>> One situation that is a bit weird after the split is that the
>>>> original dynlink (that was including both what is now dynlink and a
>>>> previous version of the conditional attributes specification) came
>>>> out of an individual draft that has IPR disclosures on it:
>>>
>>> Up to and including IPR disclosure #2508 on 2014-12-19, disclosures
>>> against I-Ds specified the draft version number. For reasons that I
>>> have forgotten, if I ever knew them, disclosures since then have not
>>> specified the version number. That change creates the breakage that
>>> you describe, and I believe it's a *serious* bug - disclosures must
>>> be against a specific version number rather than against unspecified
>>> past and future versions of the draft.
>>>
>>> RFC 3668 (applicable from February 2004) and RFC 4879 (applicable
>>> from March 2005) were rather precise:
>>>
>>> "The disclosure must also list the specific IETF or RFC Editor
>>> Document(s) or activity affected.  If the IETF Document is an
>>> Internet-Draft, it must be referenced by specific version number."
>>>
>>> RFC 8179 (applicable since May 2017) is equally precise:
>>>
>>> "An IPR disclosure must include ... (c) the specific IETF Document(s)
>>> or activity affected, and (d) if the IETF Document is an
>>> Internet-Draft, its specific version number."
>>>
>>> I really hate to say this, but it seems to me that all IPR
>>> disclosures against I-Ds since 2014-12-27 do not conform to RFC 4879
>>> or RFC 8179. Indeed, RFC 8179 confirms that the change made in late
>>> 2014 was a serious error.
>>>
>>> If the disclosure that Carsten refers to had been correct, i.e.
>>> specified the version number as required, his problem would not
>>> exactly go away, but it would be tractable.
>>>
>>> Does anyone here know why we started to break our own rule in
>>> December 2014?
>>>
>>> How can we fix the ~3500 broken IPR disclosures since then?
>>> Presumably, we must deem them to apply to the draft version that was
>>> current at the date of the disclosure.
>>>
>>>        Brian
>>>
>>>>
>>>> https://datatracker.ietf.org/ipr/search/?submit=draft&id=draft-ietf-core-dynlink
>>>>
>>>>
>>>> IANAL, but it very much seems the disclosure might be related to the
>>>> part of the specification that went into conditional-attributes.
>>>> So, after the split, these disclosures are now associated in our
>>>> database with the wrong draft.
>>>>
>>>> The then holder of the patent claims in the disclosures apparently
>>>> has since decided to abandon those claims.  It does not seem easy to
>>>> obtain a formal statement from them to that effect (the abandonment,
>>>> however, appears to be easy to ascertain from the patent databases;
>>>> IANAL though); after all, there no longer is any patent claim that
>>>> needs to be disclosed.
>>>>
>>>> So it seems I have two questions:
>>>>
>>>> (1) How do we handle IPR disclosures that stick to the wrong draft
>>>> after a reorganization like the split we did?
>>>>
>>>> (2) How do we handle the absence of negative IPR disclosures?
>>>>
>>>> Grüße, Carsten
>>>>
>>>>
>>>> (This could also be discussed in the more general context of
>>>> updating IPR disclosures, e.g.,
>>>> <https://mailarchive.ietf.org/arch/msg/ipr-wg/IHlcUZmeyanJ1f_hXKqu9yvSy_8>.)
>>>>
>>>>
>>>>
>>>> _______________________________________________
>>>> Ipr-wg mailing list
>>>> [e-mail]
>>>> https://www.ietf.org/mailman/listinfo/ipr-wg
>> _______________________________________________
>> Ipr-wg mailing list
>> [e-mail]
>> https://www.ietf.org/mailman/listinfo/ipr-wg