[rfc-i] "canonical" URI for RFCs, BCPs

Paul Hoffman paul.hoffman at vpnc.org
Tue Jan 26 14:03:58 PST 2010

At 1:47 PM -0800 1/26/10, Joe Touch wrote:
>Paul Hoffman wrote:
>> At 3:47 PM -0500 1/26/10, Henning Schulzrinne wrote:
>>> DOIs would be nice, given that every other technical paper of
>>> relevance now has one.
>> Having a DOI is not the same as a URI that is guaranteed to be
>> resolvable in the way that the controlling organization wants.
>> At 1:18 PM -0800 1/26/10, Joe Touch wrote:
>>> AFAICT, the following pattern ought to be the one with the
>>> long-term commitment:  http://www.rfc-editor.org/rfc/rfc[0-9]+.txt
>> Please, no. That means that the format can never be anything other
>> than a text file.
>Although I agree, note that this responds with a page with information
>ABOUT the document, not with the document itself (and it's not even in

Nor should it be in XML. It should be in a format that easily parsed by *everyone* resolving the canonical URL. At this point, that would be plain text and early versions of HTML. In a decade, XML might be fine.

>Until non-txt RFCs are canonical, the above is correct. That fact is
>actually stated on the page these resolve to, e.g., for 793:
>  Canonical URL:
>    http://www.rfc-editor.org/rfc/rfc793.txt
>When that policy changes, then it would be useful to provide the URL to
>the specific canonical document (.ps, .pdf, .xml, .wahoo, or whatever).

This says "because now we have text as the canonical format, we should hard-code that into canonical URLs that will be used in the future". The Web community (that is, not the IETF) have found this type of statement to be short-sighted and, more importantly, unnecessary.

--Paul Hoffman, Director
--VPN Consortium

More information about the rfc-interest mailing list