[rfc-i] Proper status for pre-IETF RFCs currently with "unknown"

"Martin J. Dürst" duerst at it.aoyama.ac.jp
Sun Nov 6 19:38:29 PST 2011

On 2011/11/07 8:45, Joe Touch wrote:
> On 11/6/2011 9:31 AM, John C Klensin wrote:
> ...
>>> My view is that history shouldn't be rewritten.
>> I guess my view is that the series is active and useful, not an
>> ossified fossil. One could claim that history was rewritten
>> when categories were assigned in the first place, when IETF
>> documents (and IENs) were merged into and replaced by the
>> series, and so on. So, if you have some particular point at
>> which you'd like to see change stop --whether you call that
>> "rewriting history" or not-- I'd like to understand that, and
>> the logic behind it, better.
> I call tacking names and categories onto published documents revisionist
> history.

Don't we do that all the time? Older stuff is updated or obsoleted by 
new stuff, and so on.

> Once published, a doc should not be modified. We understand that with
> -bis docs (i.e., we don't rewrite an RFC with the same number); why
> don't we understand that for labels?

Like Bob and others, I agree that we leave the documents themselves 
alone. But what's wrong with changing labels? Indeed, changing labels is 
extremely important for our work (defining standards). What may at one 
time have been the latest and greatest specification, the 'greatest 
thing since sliced bread', may now be outdated, useless, or whatever you 
call it.

>> FWIW, if you don't think the IESG/IETF has the authority to
>> change names within the IETF Stream, then you should rush out
>> and appeal the recent decision to move from there Standards
>> Track levels to two and rename the second.
> Within the IETF stream is a different matter. The docs we're talking
> about predate that.

And I think it makes perfect sense to apply the "stream" concept to 
stuff before the IETF.

As a totally arbitrary example (and with apologies to Dave), let's look 
at http://tools.ietf.org/html/rfc822. Not only that RFC, but that actual 
Web page.

This is just an aside, but I want to point out that 
tools.ietf.org/html/... gives both the actually published document, 
without any changes except for added convenience by adding links, and 
the metainformation (outside labels) at the top.

Getting back to the main point: This document has a label of "STANDARD". 
But it's obsolete. For everybody except those really, really familiar 
with the situation, this means: "Go figure.".

> My second concern is the IETF attaching labels to docs they didn't
> generate. That's outside their scope IMO.

Well, RFC 822 dates from 1982, and the first IETF was in 1986, but the 
IETF was formed, I assume (I wasn't around then), because a more 
organized form of moving internet standards forward was deemed 
appropriate. In simple words, the IETF "took over".

What you are saying reads to me as if you were claiming that the US 
Congress cannot take a (British) law from before 1776 off the books (for 
the US) because it didn't exist at that time. I'd guess everybody in the 
US would disagree with you.

So I'd suggest:

The IETF should take all the authority it needs to reclassify, where 
deemed necessary, any specification that in an area that they are now 
dealing with. That will help everybody understand e.g. that they'd be 
better informed looking at RFC 5322 than at RFC 822. It has nothing to 
do with rewriting history, but everything with doing the job we are 
supposed to do.

For the other kinds of documents (e.g. things closer to just research 
proposals, reports,...), I'd personally remove the "unknown" label and 
leave it as unlabeled. I also think we don't have to label everything in 
great detail. For much stuff, it doesn't make any difference at all.

Regards,    Martin.

More information about the rfc-interest mailing list