Re: Working documents repository

[Vernon Schryver <vjs at calcite.rhyolite.com>]

> From: Lloyd Wood <L.Wood at surrey.ac.uk>

> > Old I-D's could be available to individual historians and lawyers for
> > limited times (e.g. 6 weeks) and after application to the IESG for a
> > username and password.  
>
> Your suggestion goes against the-open-inclusiveness-spirit-of-the-IETF.
>
> What special qualifications/rights/level of cluefulness do historians
> and lawyers have that permit this for them? Would you limit demand
> based on, I don't know, ability to pay as clear demonstration of need,
> say?

I did not mean to imply that the IESG would check any credentials.  The
idea was to make the old I-D's available to anyone who really wanted to
see them, but without publishing them, and so also meet the objections of
those who don't want to see the old versions on-line and with the same de
facto imprimatur as a standards track RFC.

As much as none of us like the policies on on-line standards of the ITU,
ANSI, IEEE, etc., their policies do ensure that the only bad standards
from those bodies really are.



> (If available infrastructure can't stretch to mailing list archives I
>  can't see it stretching to something as fiendishly complex as a
>  password system.)

"fiendishly complex"?  Do you use the same Internet and WWW, where lots
of stuff such as free newspapers require a username and password? 


> > That would keep J. Random Marketoon from claiming compliance with an
> > I-D, and not mentioning that it is #3 in the archive on venera
> > instead of the #97 in the active directory.
>
> Hardly. The people most likely to believe the claim would never check;
> the people who would check would know where to look.

Then you think the old I-D's should be on line and available.
Please argue with those who strongly disagrees.  I don't.


> ...
> Actually, having repositories of expired I-D's is arguably symptomatic
> of abuse of the standards process. Linking to same is arguably
> symptomatic of encouraging abuse of the standards process; if you
> really thought the ideas in the draft were any good you'd have revised
> the draft in accordance with review comments and resubmitted it. If
> you thought your ideas were still good right now you'd be able to
> point to a current internet-draft with your name on as evidence of
> this. 

Things in politics or engineering, are never as simple as that.
On the way to producing a standard (or any design), you always throw away
many good ideas, or the standard (or any design) is junk.

Keeping the old I-D's out of sight does keep the marketoons from
making bogus claims.  Marketoons don't make their bogus claims because
they are cunning, but because they found something that fits the
product and sounds good.  What they can't find, they can't quote.

To be honest and politically incorrect, the problem is not marketoons,
but people who claim to have technical clues and know how the IETF works,
but don't.


> Persistence should lead to review with an explicit view to
> publishing as an informational RFC - but this does lead to more
> informational RFCs.

Are you saying that all I-D's should be published as Informational RFC's?
If not, from where do those reviews come?


> IMO forcibly beginning every RFC (and not just standards-track 'see
> STD1' ones) with a boilerplate indicating it might be
> obsoleted/updated and to see the index/summaries would IMO be a good
> idea, since it helps prevents RFCs being read out of context.

HAH!  Such boilerplates make the lawyers happy, but have no effect in the
real world.  Nothing keeps anything from being read out of context.


>                                                               We
> already put this on drafts with good reason:
>
>     Internet-Drafts are draft documents valid for a maximum of six
> ...

That is an unreasonably optimistic view of that or any boilerplate.  At
best, it answers those who demand to know why an expired I-D's is no longer
on line.  It certainly does not keep anyone from refering to I-D's in
their documents.  Like most boilerplate, most people never even see it.
Somewhere soon after learning to read, everyone learns to not see shrink
wrap licenses, safety warnings, and legalistic boilerplate mumbojumbo.



> ...
> We need to stamp something like:
>
>     This RFC document may be updated, replaced or obsoleted by later
>     RFC documents at any time. It is inappropriate to cite this
>     RFC document as a reference without citing its status as
>     {Informational|Best Current Practice|whatever} and checking
>     the current rfc-index for this RFC's current status.
>
> Yes, it's blatantly obvious - to you. In a world where unpublished
> material is widely available, clarity of implicit context is
> increasingly important, and explicitly distinguishing the class and
> validity of an RFC can't hurt.

It would not matter if you required every reader of every I-D to repeat
such a disclaimer in their own words before giving them the rest of the
text.  It would still have no effect in the real world, and few significant
effects in unreal worlds such as the trade rags and courts.


Vernon Schryver    vjs at rhyolite.com