Re: [rfc-i] Public archival of AUTH48 communications

[Ted Hardie <ted.ietf@xxxxxxxxx>]

On Fri, Feb 25, 2022 at 4:59 PM Eric Rescorla <ekr@xxxxxxxx> wrote:

On Fri, Feb 25, 2022 at 7:43 AM Francesca Palombini <francesca.palombini@xxxxxxxxxxxx> wrote:

Hi all,

 

I am monitoring the discussion, listening to the feedback, and hope to answer to all comments later on.

 

Ted: Just a clarification question. As a separate mailing list was also the proposal, I am trying to understand the difference with Ekr’s suggestion.

If I understand correctly, the difference with what Ekr is suggesting is that there would be one AUTH48 mailing list for each AUTH48 conversation, is that the idea? This was discussed, but the same sort of subscribing and filtering can be achieved locally even with one single list, and would require less tooling and RPC change, so I am not sure I see the interest of setting it up that way.

No, I'm suggesting that it doesn't go to the main WG mailing list. I am indifferent to whether it is a single list for all AUTH48 communications for a WG or one for each RFC.

What I don't think we should do is just change the setup for each WG mailing list so that suddenly everyone gets the AUTH48 stuff whether they care or not, thus forcing most of those people to set up filters to filter it out.

I see I misrepresented you in my mail to Francesca.  I thought you were proposing to do it by WG, rather than by draft.  To me, that's the natural unit of review here.  That's why I initially suggested using the WG mailing list, and why I would be fine with setting up a per WG AUTH48 list as a separate list.

That is, in any case, what I now suggest.

regards,

Ted Hardie

 

-Ekr

 

 

Thanks,
Francesca

 

From: iesg <iesg-bounces@xxxxxxxx> on behalf of Ted Hardie <ted.ietf@xxxxxxxxx>
Date: Friday, 25 February 2022 at 16:15
To: Eric Rescorla <ekr@xxxxxxxx>
Cc: Working Group Chairs <wgchairs@xxxxxxxx>, admin-discuss@xxxxxxxx <admin-discuss@xxxxxxxx>, IETF <ietf@xxxxxxxx>, RFC Interest <rfc-interest@xxxxxxxxxxxxxx>, IESG <iesg@xxxxxxxx>, irtf-announce@xxxxxxxx <irtf-announce@xxxxxxxx>, IETF Announcement List <ietf-announce@xxxxxxxx>
Subject: Re: [rfc-i] Public archival of AUTH48 communications

On Fri, Feb 25, 2022 at 2:19 PM Eric Rescorla <ekr@xxxxxxxx> wrote:
>
> While I think transparency is good, I think sending them through the WG list would be a mistake.
>
> In many cases there is a tremendous amount of back and forth about small details that I would prefer not to be bothered with for every WG list I am on.
>
> I don't object to a separate list that WG members could subscribe to.
>
> -Ekr
>
This would work for me, and I think it is a nice parallel to the way
the GH comment streams are set up in the working groups that use them.
There is a separate list you can subscribe to, and it's your choice as
to whether to intermingle that in your folders with the main WG list.
I personally keep them nested under the main list, and I'd do the same
here, but I like the flexibility this proposal offers.

regards,

Ted Hardie


>
> On Fri, Feb 25, 2022 at 12:51 AM Ted Hardie <ted.ietf@xxxxxxxxx> wrote:
>>
>> Thanks for considering the transparency question seriously.  I
>> believe, however, that you are missing a more obvious implementation:
>> cc'ing the working group mailing lists for AUTH48 discussions.  That
>> keeps the information related to a single document in a common place,
>> which makes it easier to actually follow the threads of work than it
>> would be if it were in a dedicated list.  While your proposal is
>> publicly archived, that approach would require any interested
>> individual to grovel through the archives periodically to see if there
>> were anything of interest, which makes it easy to miss the discussions
>> and more difficult to link back to the working group discussions.
>>
>> You may be concerned that this approach would lead to the working
>> groups' members jumping into the last call discussions, since those
>> lists are not read only.  You can avoid them doing so directly by
>> channeling the AUTH48 mail to the mailing through a send-only address
>> which does not receive mail.*  That would not stop a mailing list
>> member from reaching out to one of the parties individually, but your
>> proposed implementation doesn't stop that either.
>>
>> If an IETF document does not have a working group (as an AD-sponsored
>> document might not), then the mail could go to the area list.
>>
>> Since this approach would be specific to the IETF, it would allow
>> later tweaks to be done directly by the IESG, without necessarily
>> coordinating them with other streams which might have different needs.
>>
>> regards,
>>
>> Ted Hardie
>>
>> *Passing them through a send-only address also potentially gives you a
>> step in which to process them so as to elide specific information that
>> you might not want to share early, such as the RFC number or links to
>> the RFCs-to-be.  If you do not choose this implementation, you may
>> still want to consider whether those details belong in the archived
>> list, since references to the RFC number may otherwise appear in
>> advance of actual publication.  Since some AUTH48 issues take a fair
>> amount of time to resolve, that can actually result in a fair amount
>> of cruft.
>>
>> On Thu, Feb 24, 2022 at 10:25 PM The IESG <iesg@xxxxxxxx> wrote:
>> >
>> > The IETF prides itself on its open process and transparent
>> > communication, with one of our core principles being our commitment to
>> > making all materials related to our standards process and other
>> > activities publicly available.
>> >
>> > For a document undergoing publication within the IETF Stream, most of
>> > its history can be traced by exploring the mail archive - from the first
>> > submission, to Last Call comments, to IESG evaluation. However, once the
>> > document enters the RFC Editor queue, the communication between authors
>> > and RFC Production Center (RPC) is only visible to a specific set of
>> > people: RFC editors, authors, the responsible ADs and WG chairs (when
>> > applicable).
>> >
>> > In order to increase transparency during the final stages of document
>> > editing, the IESG, IAB, ISE, IRSG, Temporary RFC Series Project Manager,
>> > RPC, and Tools team are considering allowing anyone to search and read
>> > AUTH48 conversations about specific documents (or clusters of
>> > documents). It is not a goal to actively involve more people in the
>> > conversation between authors and RFC editors.
>> >
>> > Our proposal: to set up a public mailing list of AUTH48 conversations,
>> > for archival purposes only, i.e., read-only. The RFC editors, when
>> > initiating the conversation with the authors will CC this mailing list.
>> > All further responses usually maintain the CC addresses, and as a
>> > consequence will be archived in the mailing list archive.
>> >
>> > All AUTH48 discussions of all documents will be archived on the same
>> > mailing list. Searches and filtering will be available based on the
>> > mails’ content and metadata, including draft name, RFC-to-be number,
>> > cluster number, sender, and date.
>> >
>> > Opt out: the authors and RFC editors are able to opt out from archival,
>> > by removing the mailing list from the CC. Although we don’t envision
>> > that this will be necessary often, there may be cases where sensitive
>> > information needs to be shared.
>> >
>> > The initial AUTH48 mail from the RFC Editor will also include text about
>> > public archival, to make sure authors are aware. With this announcement,
>> > the IESG wants to inform the community and read feedback before we set
>> > anything up. Please send any comment to iesg@xxxxxxxx or reply to this
>> > thread by 24 March 2022.
>> >
>> > Francesca Palombini for the IESG
>> >
>> > _______________________________________________
>> > IETF-Announce mailing list
>> > IETF-Announce@xxxxxxxx
>> > https://www.ietf.org/mailman/listinfo/ietf-announce
>> _______________________________________________
>> rfc-interest mailing list
>> rfc-interest@xxxxxxxxxxxxxx
>> https://www.rfc-editor.org/mailman/listinfo/rfc-interest