Re: "Normative"? (Re: [newtrk] SRD - normative or not?)

[Douglas Otis <[e-mail]>]

On Oct 20, 2005, at 3:45 PM, John C Klensin wrote:
>
> To the extent to which, for new or updated documents at least,
> the ISDs _are_ the standard and modifications to the standard
> cannot be last-called without it, there is no plausible way for
> such a gap to develop.   Of course, imposition of  that
> requirement could either slow down or speed up document
> processing: there are plausible opinions on both sides of that
> issue.

Increasing the number of steps in front of those making changes will  
not improve upon the state of the documentation.  Avenues may be used  
that impose fewer steps.  It would be a poor outcome for RFCs to  
devolve into disembodied documents to avoid an onerous association  
with an ISD.  Conversely would the depreciated WG RFC draft process  
be avoided for the ISD draft process?  Ending up with many  
disassociated documents or massive all encompassing documents will  
not make change easier, nor will it make the review process easier.   
This is a problem we all face, and not just the IESG.  It seems  
better to breakdown the effort into smaller pieces, as it exists  
today.  The component that is missing is the cover page that ties  
them together.


>>  Allowing
>> modifications to RFCs by way of the  ISD will cause reticence
>> when making subsequent changes, especially  when new changes
>> are included or perhaps deleted.
>>
>
> Here, we disagree, perhaps based on different experiences
> getting documents through the IETF, attempting to manage WGs,
> and seeing how the IESG functions from both the inside and
> outside.  I would contend that there is already "reticence" to
> initiate revisions of RFCs for other than major issues: there is
> a large amount of work involved, a very-high-threshold approval
> process, and so on.

I agree.  Something that looks more like wiki page for the various  
endeavours would be helpful.  Think of this as a community appraisal  
mechanism that offers lower thresholds for participation than that of  
an RFC.  When people get frustrated looking at too many corrections  
and stipulations, this may provide both motivation and guidance to  
those that are likely not the originator of the initial documents.   
Lowering the thresholds in general will introduce a greater amount of  
documentation.  The goal would then be to clarify what documents  
offer value.  I suspect that the community could offer such opinions,  
if there was an available mechanism.  The SRD offers a way to  
consolidate this information with stable references.  In my view, the  
SRD should not carry the information, as this information could be  
ephemeral.  The SRD should simply serve as a reference for where to  
find such added information.

There is the annotation element within the reference element that  
could allow information to be posted with the SRD.  The information  
accumulated within the annotation element would be a separate process  
from that of the SRD itself.

>> In my view, the  SRD would
>> have a very low profile to ensure establishing the RFC
>> relationships will not burden the process.  However, this has
>> not  been a decision made by the WG.
>>
>
> Just my opinion but the difficulty that arises, and the point at
> which that nasty word "normative" applies, is that the mere
> decision about what documents go together and which ones do not
> is either a substantive action of some importance or it is a
> personal opinion of some author or group of authors.

It could be either way.  I see a large number of SRDs only used to  
keep track of ongoing work by various WGs.  Every member of the WG  
could publish their own view of how things go together.


> If is is a substantive action, then it needs, IMO, IESG review and  
> Last Call about the decision itself, whether there is non-link text  
> in the SRD or not and we are well out onto the slippery slope  
> toward ISDs.

I think that the Last Call of the related I-D should include the SRD  
as the "cover-page."  This "cover page" could reference several  
outstanding I-Ds and more than one SRD could reference the I-D.  When  
an I-D becomes approved along with its "cover-page(s)", then the base  
SRD is updated where the RFC reference replaces the I-D reference.   
Any outstanding I-Ds would remain in the WIP-SRD that would have a  
zeroed iteration-number, but an increased base SRD serial-number.

>   If it is a personal opinion, then your observations above about  
> named authors would seem to apply.  It may be my lack of  
> imagination, but I don't see how one can have it both ways.

Retaining a minimalistic approach where I-D synchronicity can be  
assured, the event of the I-D approval serves as this reference.   
Previous SRDs should remain and made available to allow an  
investigation of the history of the document and to indicate who was  
involved when questions arise.  A simple log of the I-D/SRD process  
provides all the needed information.


>> I still see the SRD as a cover page for I-Ds,  that become  
>> approved when the I-D is accepted.  The end result is  simply a  
>> stable reference.  Let the stable reference act as a means to  
>> consolidate where ancillary information is located.
>
> If it were really a "cover page of I-Ds", then why not incorporate  
> it into the I-D and be done with it? That would require no  
> procedural changes at all.

I put together a more complex WIP-SRD example of the ongoing effort  
related to SCTP.  From this example, it should be more obvious why a  
separate document is simpler.  There are already citations within  
each I-D at the time of its approval.  The desire is to _not_ revisit  
information contained in each separate RFC, but to provide a more  
dynamic linking document requiring little effort to maintain.


> If a group of I-Ds went together, why not produce an extra one that  
> was just a collection of pointers to the others, plus the usual  
> boilerplate, then ask the IESG to process the lot together.

Again, the goal was to achieve a stable reference that locates  
documents currently associated with a specific endeavour. Your  
suggestion does not seem to achieve that goal.


> That, it would seem to me, would make an easy experiment to gain
> experience with that particular view of things, again requiring
> no WG action or procedural changes.

The introduction of WIP-SRD would not require any WG or IESG action,  
where each person could provide their own namespace for that matter.   
Organize them with links and make them available and see what people  
think.  Over time, as a greater number of RFCs are incorporated  
within SRDs, then the IESG could decide whether they want to see  
these documents together with submitted I-Ds.

-Doug


.
newtrk resources:_____________________________________________________
web user interface: http://darkwing.uoregon.edu/~llynch/newtrk.html
mhonarc archive: http://darkwing.uoregon.edu/~llynch/newtrk/index.html