Re: Rule of removing adopted work (was Re: Question about pre-meeting document posting deadlines for the IESG and the community)

[Keith Moore <[e-mail]>]

On 3/19/24 05:16, Adrian Farrel wrote:

> AB is right (don’t be too surprised ;-) that establishing consensus 
> and recording history is crucial.
>
> However, “the people” is ambiguous. The I-D, once adopted, belongs to 
> the WG. The charter, once approved, belongs to the IETF.
>
> All changes should be recorded properly, and it is easy to add a note 
> to the History tab in the Datatracker. And these days, it is also easy 
> to add a pointer to an email thread.
>
> FWIW, I am just removing a draft in the MPLS WG. The draft is already 
> in the publication queue with the AD, however, it appears that there 
> is no interest in fixing the bugs or implementing the content.
>
> We have had several requests for input on the WG list, raised it on 
> the agenda today, and held a formal consensus call to abandon the work.
>
Somewhat agree.   A WG can decide to abandon a document, or substitute a 
different document for one that it was previously working on.   Yes, 
this should be a formal consensus action on the part of the WG, and 
documented appropriately.

IESG can also remove a work item from a WG's charter.  Again, this 
should be a formal action documented in the data tracker.

However, I've never been sure what "WG adoption" means and I don't 
recall ever seeing a formal definition of that.

I don't think that an I-D "belongs" to a WG, or at least, I think use of 
terms of "ownership" can be misleading at best.   In terms of copyright, 
the author(s) of an I-D retain legal ownership of their original work.  
They can choose to grant (nor not grant) a license to the IETF Trust to 
make derivative works, and that permission is presumably conveyed to a 
WG if IESG includes that I-D as a work item in the WG's charter.   (The 
right to make derivative works is one of several separate rights granted 
to the author(s) of a work under copyright law and the Berne convention, 
upon creation of that work, and without the need for any formal 
registration of the work.)   As I understand the mechanism, this 
granting of a license by the author(s) to the IETF Trust is usually done 
by setting the "ipr" parameter in an xml2rfc document (which is IMO a 
horrible mechanism for that because it makes it unnecessarily cumbersome 
for authors to understand exactly what rights they're giving up).  Note 
that the license granted is "non-exclusive", i.e. the author retains 
their original rights (including the right to make derivative works) in 
addition to granting a license to the IETF Trust for a subset of those 
rights.   And the IETF Trust in turn delegates that permission to other 
IETF participants.  (I am not a lawyer, but this seems dodgy to me.)

But I have strong reservations (to put it mildly) about the idea that 
the IETF should decide to grant anything like "ownership" of a document 
to a WG or another editor, while the author(s) of the document are 
living, without at least some agreement with the author(s) about how 
changes are to be managed.   The reason is that on multiple occasions 
I've seen a document that I wrote or co-wrote, drastically changed by 
someone (it's not always been clear who) in such a way as to do serious 
technical harm to the document, including drastically compromising the 
security of the protocol described.

The original author(s) have a strong interest in keeping  their creation 
technically sound and internally consistent.  Other participants may not 
fully understand the considerations that went into the careful selection 
of text in the document, and in some cases, other participants may 
attempt to compromise the technical soundness of the document.   I don't 
believe that there is anything like adequate attention to this paid in 
the current IETF WG or RFC editing process.

So, while I understand that IETF needs "change control" over the 
standards that it produces, I'm also very concerned about the potential 
for IETF or the RFC Editor to compromise the technical soundness of 
documents, whether out of accident or malice or some combination of the 
two.   And I regard any action from IETF management or a working group 
to "take away" change control from original author(s) without at least 
having a discussion with them, as inexcusably rude at best, with 
significant potential for harm to the Internet.  Even if the IETF Trust 
technically has a legal license to make derivative works of that work.

Keith