[rfc-i] draft-iab-styleguide-02 on referencing STDs

Brian E Carpenter brian.e.carpenter at gmail.com
Sun Apr 13 13:25:56 PDT 2014


On 14/04/2014 05:03, Dave Crocker wrote:
> On 4/13/2014 9:53 AM, Julian Reschke wrote:
>> In the ~10 years that I've been working on IETF documents I have rarely
>> seen a case where citing a STD or BCP actually had an advantage.
> ...
>> Also, as one of then authors of a six-part spec that might eventually
>> become a STD, I'd really like to understand how it's going to be
>> referenced in the mid term.
> 
> 
> There were good and reasonable intentions behind creating BCP and STD as
> distinct numbering series, but they've proven to be only a hassle,
> without any actual benefit.  In the case of STD, it often mostly serves
> to create confusion (when the listing is out of sync with the latest
> version(s).)

Actually it's somewhat less problematic with BCPs, since there is only
one maturity level - therefore the current BCP is unambiguously the
union of a set of RFCs, even if some of them have obsoleted others.

In that case [BCP9999] could be a citation for the latest version
at the time of reading, whereas [RFC9998, RFC9997] could be a citation
for a specific version**.

That isn't the case for STDs, where the real standard may consist
of one or more Internet Standard RFCs amended by one or more
Proposed Standard RFCs (or legacy Draft Standard RFCs) which are
not formally at STD level. So even if you want to refer to the latest
version, [STD9999] won't work, because it won't catch any PS documents
that amend the STD. Referring to the individual RFCs is the only
solution that works.

**However, even for BCPs, things don't always work out. Consider
BCP 9. It contains material that is updated by BCP 78. So a citation
of [BCP9] leads you to the set of RFCs identified as part of BCP 9,
which does not include BCP 78. It therefore leads you to obsolete
material.

Worms, Can of, do not open.

   Brian


More information about the rfc-interest mailing list