[rfc-i] citing historic internet drafts

RFC Editor rfc-editor at rfc-editor.org
Tue Oct 21 13:58:04 PDT 2008


Part 3

----- Forwarded message from Julian Reschke <julian.reschke at gmx.de> -----

Date: Mon, 20 Oct 2008 19:00:51 +0200
From: Julian Reschke <julian.reschke at gmx.de>
To: Bob Braden <braden at ISI.EDU>
CC: rfc-editor at rfc-editor.org
Subject: Re: [rfc-i] citing historic internet drafts

Bob Braden wrote:
>
>Julian,
>
>Let me understand a couple of things.

And, first of all, many thanks for giving feedback...

>>Julian Reschke wrote:
>>>...
>>>OK, so let's have a look at the two informative ID references I'm 
>>>currently struggling with (see 
>>><http://tools.ietf.org/html/draft-reschke-webdav-search-18#section-12.2>): 
>>>
>>>    [DASLREQ]  Davis, J., Reddy, S., and J. Slein, "Requirements for DAV
>>>               Searching and Locating", February 1999, <http://
>>>               www.webdav.org/dasl/requirements/
>>>               draft-dasl-requirements-01.html>.
>>>               This is an updated version of the Internet Draft
>>>               "draft-ietf-dasl-requirements-00", but obviously never 
>>>was
>>>               submitted to the IETF.
>>>(note that this very text passed IETF last call *and* was approved by 
>>>the IESG...)
>>>Observations:
>>>- it's over 9 years old
>>>- it's an informative reference
>>>- it does not claim that it is an Internet Draft at all
>>>- it provides a URL on www.webdav.org, which is the most stable place 
>>>I can think of for WebDAV stuff
>>>- it explains in an annotation how that document actually differs 
>>>from a previous draft, but does not cite that one
>
>Why should the reader care about  the earlier Internet Draft? When I 
>read a technical
>article, or a reference from an article, I do not care about earlier 
>versions that became
>the document I can find and read; the document itself stands alone.  So 
>why not
>stop after the URL and omit the ref to an Internet Draft entirely, cited 
>or not?

Except for giving some additional information. Dunno.

So, just this:

    [DASLREQ]  Davis, J., Reddy, S., and J. Slein, "Requirements for DAV
               Searching and Locating", February 1999, <http://
               www.webdav.org/dasl/requirements/
               draft-dasl-requirements-01.html>.

would be ok?

>Then the remaining issue would seem to be the stabiity of the webdav 
>archive.
>We hate URLs like that, but in the end we sometimes have to be permissive.

I really like to understand *what* kind of URLs you dislike -- ones 
containing the original "draft-" string? That's fixable in this case (as 
I have write access to webdav.org).

>>>- publishing this as historic RFC may be possible, but requires first 
>>>a ton of updates (it has been written when the rules were different), 
>>>and would also require negotiation with the original authors
>
>Yes, well, if the IETF took their inherited culture seriously, that 
>requirements doc,
>if it is worth referencing, should have been submitted as an RFC in the 
>beginning.

Point taken.

>>>So why was this rejected by the RFC-Editor? Because it contains the 
>>>term "draft"???
>
>Actually, I am not sure, I will ask Sandy.

Good.

>>>The other one is:
>>>    [DASL]     Reddy, S., Lowry, D., Reddy, S., Henderson, R., Davis, 
>>>J.,
>>>               and A. Babich, "DAV Searching & Locating",
>>>               draft-ietf-dasl-protocol-00 (work in progress), July 
>>>1999.
>>>In this case the approved ID actually *does* use the standard format 
>>>(using "work in progress"), as I missed the problem.
>
>So, here is the problem as I see it.  An RFC is forever.  Internet 
>Drafts, according to the rules, officially vanish after 6 months.  It is 
>completely illogical for an RFC to have a reference to a document that 
>will not exist 6 months later, when the RFC may be read 10 years later.
>The "work in progress" idiom was adopted as a compromise.

Understood.

Of course with tools.ietf.org and Google internet drafts do not vanish 
anymore, and the draft name *is* a good reference.

>>>This is another case of a draft that clearly is not work in progress: 
>>>the spec that references it
>
>Wasn't it in progress in July 1999?

Yes, it was.

>>>actually is a successor of it, and it is only cited in an attempt of 
>>>explaining the history of the spec.
>>>I think the best way *for the reader* would be to state:
>>>- yes, this was an I-D, and provide the exact name,
>
>If  RFCs carry exact names of I-Ds, you are saying that the I-D 
>documents form
>a stable ("forever") document series.  Without saying that is a bad 
>idea, I believe that
>you have to follow up all
>the consequences: someone to manage the series, archive it, catalog it, 
>ensure
>uniqueness, etc.  None of that
>would be hard to do, given the current administrative machinery in the 
>IETF,
>but it has not been done.

Understood.

>>>- do not claim it is work in progress, but state that it was abandoned,
>
>This seems wrong.  It *is* abandoned today, but it was not abandoned 
>when it
>was written.

So is it allowed to both say "work in progress", and then add an 
annotation (xml2rfc <annotation> element) that gives more details?

>>>- provide a stable URL to an archived version.
>>>Such as:
>>>    [DASL]     Reddy, S., Lowry, D., Reddy, S., Henderson, R., Davis, 
>>>J.,
>>>               and A. Babich, "DAV Searching & Locating",
>>>               draft-ietf-dasl-protocol-00 (abandoned), July 1999.
>
>If you are going to include the exact I-D name (see my comments above), 
>then
>neither "abandoned" nor "work in progress" seem relevant.   There is no 
>point in
>a reference if it cannot be found, and if it can be found ... etc., etc.

Well, both are relevant to the reader; "work in progress" is plainly 
confusing because it suggests that somebody is working on the document. 
In this particular case it would mean that the development of DASL 
somehow was forked by me (leading to RFC5323), while others may be 
working on a potentially competing version...

BR, Julian

----- End forwarded message -----


More information about the rfc-interest mailing list