Terms used in rules-update-07

--On Sunday, July 30, 2006 2:09 PM -0400 "Contreras,
Jorge" 
<Jorge.Contreraswilmerhale.com> wrote:

>> Jorge has assured me in the past that any document
that's
>> approved in the IETF (i.e. by the IESG) is part of
the IETF
>> standards process, even if it isn't on the IETF
Standards
>> Track. This  doesn't apply
>> to documents from other sources (e.g. IRTF or
Independent
>> Submissions.)
>
> That's correct.

It seems to me that, if we are going to start relying on
that 
distinction in any substantive way, we are going to need to
make 
it very clear which documents are which.  At present, as far
as 
the RFCs themselves, there is no textual distinction between

IESG-approved Informational documents and independent 
submissions... unless one knows what the exact distinctions
are 
in boilerplate in a given month and then does the detective 
work.

Worse, there is a subtle interaction between this question
and 
the discussions about the RFC Editor RFP, the status of 
independent submissions, etc.  To briefly repeat what has
been 
said elsewhere, our tradition -- I thought firmly
established 
when the current management model was put in place -- is
that 
the IESG does not get to determine what the community
thinks: 
the role of the IESG is steering, management, and
interpretation 
of community consensus and conclusions after the community
has 
been asked.  The text in 2026 reinforces this: the idea of
an 
"AD-sponsored" informational or experimental
document that does 
not originate as a WG product is entirely the invention of
the 
IESG: there is no place in, or authority for, such documents
in 
our formal procedures.  If 2026 is to be taken seriously,
the 
_only_ RFCs that are IETF products are standards track
either 
standards track or products of WG efforts.

Remember that, under the "AD-sponsored" theory
of informational 
or experimental RFCs, the IESG is not required to manage a
Last 
Call on those documents, announce the action it is
considering 
taking, or involve the IETF community (or seek its
consensus) in 
any way beyond the simple fact of publication as an 
Internet-Draft.  The AD sponsorship model has seemed
relatively 
harmless to me, and possibly beneficial, but, if those
documents 
are part of the standards track, we have a whole series of 
issues about announcements, possibility for appeal, and
other 
appropriate process that, it seems to me, require
modifications 
to 2026, approval of the ISOC Board and insurance carriers,
etc.

I don't see how "IESG approved" can be used to
substitute for 
"IETF produced" unless we are going to go back
and modify 2026 
to extend the IESG's authority to decide things without any

contact with the community other than the wisdom and
intuitions 
of the IESG membership.  Jorge, it would be different if you

said "any document that went through IETF Last Call
and was 
subsequently approved" was an IETF document.  But I
just don't 
think there is and foundation for equating "IESG
approved" with 
"approved in the IETF" and hence part of the
standards process. 
It actually seems to me that such an interpretation is
fraught 
with dangers.

      john


_______________________________________________
Ipr-wg mailing list
Ipr-wgietf.org
https:/
/www1.ietf.org/mailman/listinfo/ipr-wg

> I don't see how "IESG approved" can be
used to substitute for "IETF 
> produced" unless we are going to go back and
modify 2026 to extend the 
> IESG's authority to decide things without any contact
with the community 
> other than the wisdom and intuitions of the IESG
membership.  Jorge, it 
> would be different if you said "any document that
went through IETF Last 
> Call and was subsequently approved" was an IETF
document.  But I just 
> don't think there is and foundation for equating
"IESG approved" with 
> "approved in the IETF" and hence part of
the standards process. It 
> actually seems to me that such an interpretation is
fraught with dangers.

Frankly I think this is splitting hairs. IESG approval is
the last stage
prior to RFC publication for documents originating in the
IETF (barring
a successful appeal). So it's the externally visible sign
that the IETF
process has completed for a given document.

    Brian

_______________________________________________
Ipr-wg mailing list
Ipr-wgietf.org
https:/
/www1.ietf.org/mailman/listinfo/ipr-wg

John C Klensin wrote:
>
> Remember that, under the "AD-sponsored"
theory of informational or 
> experimental RFCs, the IESG is not required to manage a
Last Call on 
> those documents, announce the action it is considering
taking, or 
> involve the IETF community (or seek its consensus) in
any way beyond 
> the simple fact of publication as an Internet-Draft. 
The AD 
> sponsorship model has seemed relatively harmless to me,
and possibly 
> beneficial, but, if those documents are part of the
standards track, 
> we have a whole series of issues about announcements,
possibility for 
> appeal, and other appropriate process that, it seems to
me, require 
> modifications to 2026, approval of the ISOC Board and
insurance 
> carriers, etc.
>
> I don't see how "IESG approved" can be
used to substitute for "IETF 
> produced" unless we are going to go back and
modify 2026 to extend the 
> IESG's authority to decide things without any contact
with the 
> community other than the wisdom and intuitions of the
IESG 
> membership.  Jorge, it would be different if you said
"any document 
> that went through IETF Last Call and was subsequently
approved" was an 
> IETF document.  But I just don't think there is and
foundation for 
> equating "IESG approved" with
"approved in the IETF" and hence part of 
> the standards process. It actually seems to me that
such an 
> interpretation is fraught with dangers. 
John,

from my perspective, your argument says that it's
impossible for the 
IESG to consider the opinions of the community on a document
without a 
Last Call.

I don't buy that.

Stuff is being produced in the IETF *all the time* through
mechanisms 
that may or may not include a Last Call, and may or may not
include WG 
processing.

Limiting the "product of the IETF" label to
stuff that is produced by 
exactly two mechanisms - IETF Last Call and IETF WG
consensus call - 
will, to my mind, be harmful.
And if we do not want to declare and police such a
restrictive label, I 
don't see any reason to declare "AD sponsored"
documents out of scope.

                                Harald


_______________________________________________
Ipr-wg mailing list
Ipr-wgietf.org
https:/
/www1.ietf.org/mailman/listinfo/ipr-wg

--On Monday, 31 July, 2006 10:57 +0200 Harald Alvestrand
<haraldalvestrand.no> wrote:

> John C Klensin wrote:
>> 
>> Remember that, under the "AD-sponsored"
theory of
>> informational or  experimental RFCs, the IESG is
not required
>> to manage a Last Call on  those documents, announce
the
>> action it is considering taking, or  involve the
IETF
>>...
>> I don't see how "IESG approved" can be
used to substitute for
>> "IETF  produced" unless we are going to
go back and modify
>> 2026 to extend the  IESG's authority to decide
things without
>> any contact with the  community other than the
wisdom and
>> intuitions of the IESG  membership.  Jorge, it
would be
>>...

> John,
> 
> from my perspective, your argument says that it's
impossible
> for the IESG to consider the opinions of the community
on a
> document without a Last Call.
> 
> I don't buy that.
> 
> Stuff is being produced in the IETF *all the time*
through
> mechanisms that may or may not include a Last Call, and
may or
> may not include WG processing.
> 
> Limiting the "product of the IETF" label to
stuff that is
> produced by exactly two mechanisms - IETF Last Call and
IETF
> WG consensus call - will, to my mind, be harmful.
> And if we do not want to declare and police such a
restrictive
> label, I don't see any reason to declare "AD
sponsored"
> documents out of scope.

Harald,

First, Jorge's clarification takes this particular debate
outside the scope of this conversation/ WG so we don't need
to
settle it to make a decision on rules-update-07 (although I
think it may indicate that document needs some clarification
and
that, in turn, might argue against its approval at this
point).

As to the more general topic, if the IETF is going to
preserve
and enhance its reputation as a body that produces consensus
standards based on solid, community-based engineering, it
seems
to me that documents cannot be produced as "IETF
approved" and
statements cannot be made that include equivalents of
"the IETF
says" or "the IETF believes" without at
least some serious
attempt at giving the community a chance to review and
provide
input into a decision (without having to rely on appeals as
a
review mechanism).  It is clear to me that the path 

	* someone produces an I-D
	* the author gets an AD, perhaps an AD with whom the
	author has a special relationship such as employment by
	the same company or working on similar projects, to
	"sponsor" the document.
	* the IESG reviews the document internally, with no
	announcement to the community that it is doing so (I
	don't consider inclusion in an IESG agenda under
	"individual submissions via AD" to be such an
	announcement)
	* the IESG passes the document to the RFC Editor with a
	requirement that it be published.
	* no formal announcement is made to the community that
	the document was under consideration, or approved, until
	the RFC is published.

does not meet the test of a serious attempt at giving the
community a chance to review and provide input.   Now,
contrast
this with a WG output in which there is a charter that lists
the
relevant document as an output.  There is at least notice to
the
community that the work is being undertaken.  There is
potential
for discussion within the WG and most such documents are
actually discussed.  One would hope that, if there is
significant dissent --or significant indifference-- within
the
WG, someone would notice that and at least raise it with the
AD,
perhaps forcefully.  So claiming that such a document is an
IETF
product is at least plausible.

There are lots of cases in between.  Indeed, I believe that
"on
the WG charter, produced as a WG-identified document,
mentioned
in an IETF meeting or two, but never really discussed in the
WG"
would fall in between.   If you have a model for
differentiating
those from developmentand passage through the IESG without
community notice but without relying on IETF or WG Last
Call,
please tell us, preferably in the form of an I-D that
modifies
2026, what that mechanism might be.

FWIW, I see at least one way that the above goals could be
accomplished without requiring that the IESG issue and
monitor
Last Calls.  The RFC Editor could, in principle, post
notices
that documents that were not the result of IETF (or perhaps
WG)
Last Call were under final review for RFC publication and
allowing a short window for comments.  Some of us believe
that
should be done for independent submissions anyway, so that
would
be a small extension to that idea.  Of course, that would
require giving the RFC Editor the authority to either reject
an
AD-Sponsored document or to bounce it back to the IESG and
insist on a Last Call if there were significant dissent...
but
it would at least avoid adding to the burden of Last Calls
for
non-controversial cases, and would not, I believe, require
any
modifications to 2026.

     john


_______________________________________________
Ipr-wg mailing list
Ipr-wgietf.org
https:/
/www1.ietf.org/mailman/listinfo/ipr-wg

At 9:46 AM -0400 7/31/06, John C Klensin wrote:
>  It is clear to me that the path
>
>	* someone produces an I-D
>	* the author gets an AD, perhaps an AD with whom the
>	author has a special relationship such as employment by
>	the same company or working on similar projects, to
>	"sponsor" the document.
>	* the IESG reviews the document internally, with no
>	announcement to the community that it is doing so (I
>	don't consider inclusion in an IESG agenda under
>	"individual submissions via AD" to be such
an
>	announcement)
>	* the IESG passes the document to the RFC Editor with a
>	requirement that it be published.
>	* no formal announcement is made to the community that
>	the document was under consideration, or approved,
until
>	the RFC is published.

Well, first let me say that ADs who sponsor documents are
already
concerned about perceived conflict of interest, so your
second bullet
is pretty unlikely for "employment by the same
company" unless
their relationship within the company is quite distant and
the topic
one where AD sponsorship is common (Even in some of those
cases,
co-ADs are asked to consider the document instead).

And that raises my second issue:  we have lots of history
that says
keeping working groups around forever creates problems, so
there
are topic areas where the IETF work product is the
foundation of
the industry use of a protocol but there is no active WG. 
What an
AD should look for in community review of a document in some
of those
cases is well defined (in the URN NID case, for example,
explicit review is called
out by the urn-nid list; media type review, explicit review
is called
for by ietf-types list, etc.).  In others, you have the
mailing list of the closed
working group to go on (e.g. LDAPEXT for LDAP extensions)
and potentially
a directorate (LDAP has one such).  For standards-track
document, ADs
should (and I believe usually do) consult the relevant lists
(even if there
are no WGs) as well as putting out an IETF Last Call.

We could, I believe, explicitly require an IETF Last Call
for all Informational
and Experimental documents that are AD sponsored; at the
moment, it is
a judgement call by the AD based on what level of community
review
something needs or has already received.  If we go that
route, though, I
think we have to be pretty aware of what it means for review
cycles
and delay.   

We've wandered off topic for this WG, though, and I suggest
continuing in
private email or on ietfietf.org.
			regards,
				Ted











>does not meet the test of a serious attempt at giving
the
>community a chance to review and provide input.   Now,
contrast
>this with a WG output in which there is a charter that
lists the
>relevant document as an output.  There is at least
notice to the
>community that the work is being undertaken.  There is
potential
>for discussion within the WG and most such documents are
>actually discussed.  One would hope that, if there is
>significant dissent --or significant indifference--
within the
>WG, someone would notice that and at least raise it with
the AD,
>perhaps forcefully.  So claiming that such a document is
an IETF
>product is at least plausible.
>
>There are lots of cases in between.  Indeed, I believe
that "on
>the WG charter, produced as a WG-identified document,
mentioned
>in an IETF meeting or two, but never really discussed in
the WG"
>would fall in between.   If you have a model for
differentiating
>those from developmentand passage through the IESG
without
>community notice but without relying on IETF or WG Last
Call,
>please tell us, preferably in the form of an I-D that
modifies
>2026, what that mechanism might be.
>
>FWIW, I see at least one way that the above goals could
be
>accomplished without requiring that the IESG issue and
monitor
>Last Calls.  The RFC Editor could, in principle, post
notices
>that documents that were not the result of IETF (or
perhaps WG)
>Last Call were under final review for RFC publication
and
>allowing a short window for comments.  Some of us
believe that
>should be done for independent submissions anyway, so
that would
>be a small extension to that idea.  Of course, that
would
>require giving the RFC Editor the authority to either
reject an
>AD-Sponsored document or to bounce it back to the IESG
and
>insist on a Last Call if there were significant
dissent... but
>it would at least avoid adding to the burden of Last
Calls for
>non-controversial cases, and would not, I believe,
require any
>modifications to 2026.
>
>     john
>
>
>_______________________________________________
>Ipr-wg mailing list
>Ipr-wgietf.org
>https:/
/www1.ietf.org/mailman/listinfo/ipr-wg


_______________________________________________
Ipr-wg mailing list
Ipr-wgietf.org
https:/
/www1.ietf.org/mailman/listinfo/ipr-wg