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

Brian E Carpenter brian.e.carpenter at gmail.com
Wed Jan 27 14:38:24 PST 2010

On 2010-01-28 10:52, Julian Reschke wrote:
> Brian E Carpenter wrote:
>>  ...
>> Exactly. Formal citations of RFCs issued in ASCII will always remain
>> as citations of the ASCII version, for ever. If we ever do change to
>> .fubar as the new canonical format, that will not change anything for
>> the .txt RFCs back to RFC 1.
>> ...
> That's true, but that's not really relevant.
> Let's assume for a moment that we'll move to HTML as standard format by
> April 1, 2010.
> What is your advice for people generating URIs then? For each URI being
> generated, check with the RFC Index about when the RFC was published,
> and generate a different URI format?
> What if we decide to serve HTML versions even for earlier RFCs? Maybe
> based on the TXT version (monospaced font etc...), or even based on the
> XML2RFC source? Do we still serve that with a .TXT extension then? What
> about historic RFCs where the canonical version is *not* the TXT
> version, maybe because they were scanned in?
> The media type simply does not *belong* into the URI. Period. And if
> it's not part of the URI it does not matter when we switch, and to what
> format. The URI format will always be the same.

Yes, that would be best, but then it would need to be direct, not a URI to
a metadocument, which is where http://www.rfc-editor.org/info/rfc1
takes me.

Unfortunately http://www.rfc-editor.org/rfc1 and http://www.rfc-editor.org/rfc/rfc1
are both 404.

http://tools.ietf.org/rfc/rfc1 does the right thing (IMHO).

The article I mentioned will recommend http://www.rfc-editor.org/rfc/rfc1.txt
explicitly as a URL (not URI).


More information about the rfc-interest mailing list