Re: FW: Last Call: <draft-farrell-ft-03.txt> (A Fast-Track way to RFC with Running Code) to Experimental RFC

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



--On Tuesday, January 22, 2013 16:31 -0500 Thomas Narten
<narten@xxxxxxxxxx> wrote:

> FWIW, I share Joe's concerns. And Stephen's responses don't
> really change my mind.
> 
> This document seems to have a bit of missing the forest for the
> trees. In the overall scheme of things, I don't believe the
> draft will materially help, and is at best a distraction from
> dealing with meaningful problems.
> 
> The crux of the issue is that any attempt at "fast tracking" is
> fundamentally about short-circuiting some aspect of our review
> processes. We should only do that very carefully. In almost
> all cases, individual judgement is needed as to when it is
> appropriate to do that. ADs/WG chairs already have some
> flexibility here. e.g., a WG can skip WG LC if they think its
>...

Hi.
First of all, I am completely in agreement with Thomas and his
analysis.  Anything  that can reasonably and appropriately be
done by this sort of effort can be done by discretion without
adoption of a formal procedure and adoption of a formal
procedure creates additional and unnecessary risks. 

As co-author of the process experiment model, I also object to
its use when it is not demonstrably necessary, for example to
give extra status and force IETF-wide use where the actions
suggested can be carried out as a matter of normal discretion
(see Section 5 of the document).  In this particular case, a
better experiment would be to use a process like this for some
otherwise-relevant drafts and in some areas only and compare the
results.  The "force IETF-wide use" part doesn't apply as long
as its invocation is a matter of WG Chair and AD discretion and
belief that the right conditions are met, so, other than
consuming a rather large amount of community time since it was
first proposed almost two months ago, I just don't see what it
accomplishes other than adding to the collection of process
documents and constraining the review process, including about
issues that running code doesn't help to inform.  The process
experiment mechanism is, IMO, an entirely inappropriate way to
simply make a statement that the IETF really, really, does
believe in running code and interoperability... especially since
one would hope that such a statement would not expire after a
year.

Another thing this draft accomplishes is to shift the
responsibility for determining whether a document is "good
enough" from an IESG interpretation of community consensus to
the decision of a single AD as to whether or not changes are
required (Section 3, item 5).  While it would probably improve
speed of processing, it is not, in general, a desirable change,
especially if one believes in the first part of the Clark quote.

It is probably just a nit in the grand scheme of things but, if
the responsible AD doesn't have time to do a sufficient review
to be confident in voting "yes" (Section 3, item 7) how can she
possibly sign off on the belief than no changes are required
(Section 3, item 5)?  It seems to me that situation creates a
situation in which an AD can initiate this process, take the
word of the WG that everything is in order, and then not only
avoid responsibility by abstaining but use the provisions of
Section 3 item 8 to let the document go through (as long as some
AD is willing to say "yes") without taking real responsibility.
Especially because we don't have firm rules against a WG Chair
and responsible AD coming from the same organization, I would
hope that those who are concerned about IETF CoI and antitrust
policies would take a careful look at that scenario because it
would seem to avoid many of the protections that are inherent in
wide community review and individual responsibility.  Other than
that, it is nothing more than a specific scenario consistent
with Thomas's concern that this procedure could be used to
short-circuit effective review.

I'm not going to quote extensively from Thomas's other arguments
or waste bits trying to summarize them, but let me add a few
observations to Thomas's about the document and how it is being
handled.

> If it is really urgent to get a document done, it is far
> better to take steps to make sure the editors are engaged and
> responsive. That is more likely the real issue!!!

And that no AD is going to screw around nit-picking or delaying
things in other ways.  Based on experience, I have absolutely no
confidence that a prohibition against anything that isn't
important (such as the one in this draft) will stop that
behavior if, e.g., someone thinks the document is
incomprehensible or ignores important issues.  (And, as Thomas
points out, it shouldn't).
 
> Running code is valuable. Always has been, always will be. But
> we need to resist the temptation of making "running code" more
> equal than other criteria or putting it on a pedestal as some
> sort of holy grail. Running code by itself is just a sound
> bite. The importance of running code is what it tells us about
> a protocol specification. Just the mere fact that there is
>...

Moreover, Dave Clark and sound bites notwithstanding,
interoperability is lots more important to us than running code,
partially because, while a lot of things might be learned by a
good-faith implementation of a specification by someone not
involved in its development, an implementation by an author or
other active participant in the development of a document may
actually tell us very little without third-party and
interoperability testing ... including not telling us whether it
is actually "running", much less conforming, code.  The latter
is a _huge_ can of worms if it is going to be a procedural
gating factor in the way the document proposes.

In that light, the document says "The spirit [...] is that early
implementations and ongoing development and testing throughout
the protocol work can significantly improve the quality of a
protocol and of its specification."  I agree with that and hope
everyone else does too.  But, other than general encouragement
and observations that such testing might be appropriate (e.g.,
in Section 4 item 13) this proposal makes no provision for
enabling or encouraging "ongoing development and testing
throughout the protocol work".  Instead, it concentrates only on
an assertion that an implementation exists, that it is
conformant enough to be informative, and that it "runs".  I
would have more sympathy with this proposal if its gating
condition was serious interoperability testing among independent
implementations, but that could still be done on a discretionary
basis (as 2026 points out in Section 4.1.1).



> But overall, I see this document at best as a no-op. However,
> I fear that it can be used to short-circuit our review
> processes in a way that doesn't help the IETF and that won't
> really speed things up enough to justify the downsides.

Yes.

In addition, this document moves a number of informal IESG
procedures -- including the DISCUSS criteria and even the
Shepherd report requirements (now Informational in RFC 4858 plus
a number of less formal "statements" and "templates" that have
sometimes changed rapidly) -- into the status of formal
procedural requirements despite the details of those procedures
never having gotten community consensus about those
specifications.  Under normal circumstances, the relationship
between this document and those specifications would require
additional actions to prevent this document from going into
permanent publication hold.   [DCRIT] is called out as a
normative reference without the discussion and alerts normally
required; the shepherding procedures and templates are not
called out at all.

I'm also feeling a need to say something that people seem to be
missing or avoiding.  In my experience in the IETF, we almost
never allow a individual submission document that is obviously
very controversial to turn into a extended debate on the IETF
list with the author trying to refute every comment.   Sooner or
later (usually sooner), the sponsoring AD says "hey, not ready
for the IETF List, much less Last Call" and pushes it off into
either a separate mailing list (to emerge only when there is
real consensus) or asks the author to take steps to propose and
organize a WG.

So, Adrian, noting the ratio between discussion of this draft on
the IETF list in the last few weeks and discussions of
everything else, how long does professional courtesy to another
IESG member (presumably in combination with your thinking this
is a good idea) encourage you to allow the unintentional DoS
attack on the IETF list to continue? 

regards,
   john



[Index of Archives]     [IETF Annoucements]     [IETF]     [IP Storage]     [Yosemite News]     [Linux SCTP]     [Linux Newbies]     [Fedora Users]