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

Brian E Carpenter brian.e.carpenter at gmail.com
Fri Apr 11 13:28:56 PDT 2014


This is indeed all very problematic.

On 11/04/2014 18:57, Julian Reschke wrote:
> This text is new, and I think it has lots or problems:
> 
>> 4.8.6.3. Referencing STDs and BCPs
>>
>>
>>    Standards (STDs) and Best Current Practices (BCPs) may consist of a
>>    single RFC or multiple RFCs.  When an STD or BCP that contains
>>    multiple RFCs is referenced, the reference entry should include ALL
>>    of the RFCs comprising that subseries.  The authors should refer to
> 
> Does that imply that the reference by STD number is *mandatory*?

I don't think that is something that can be decided except by the
IETF, and as far as I know it hasn't been discussed by the IETF.

It would be a *major* change to current practice, and it's also
broken:

>>    specific RFC numbers as part of the text (not as citations) and cite

That's not OK. It's actually worse than not OK. It's *wrong*. A normative
reference must be to a specific version of the cited standard, and
the rather broken STD designation doesn't support versioning.

>>    the subseries number.  Inclusion of the URI to the STD or BCP info
>>    page is now recommended.  The text should appear as follows:
>>
>>       See RFC 1034 [STD13].

No. Because if subsequently RFC 1034 is obsoleted by RFC 10034, the
meaning of the citation changes and the technical details would
no longer be correct. STDs are not immutable. RFCs are immutable,
which is why they are the "atom" of normative citations.

Just to complicate it, RFC 1034 is updated by RFC 5936, which
is Proposed Standard. Is that part of STD13? No. Does it modify
the meaning of RFC 1034? Yes.

Look, this is a mess in the standards process, which people
tried to fix 10 years ago without success:
http://tools.ietf.org/html/draft-bradner-stdproc-isd-01 .
Also see http://tools.ietf.org/html/draft-klensin-std-numbers-01 .

This is not a topic for the Style Guide. It's an IETF process
discussion. The change needs to be backed out.

    Brian

> 
> That ruins hyperlinking in the generated output formats. Also, it'll
> make linking to specific sections and marking that information up in a
> readable way a headache.
> 
> ...augmenting a reference with the information that it is part of a STD
> is good, but pretending it *is* the reference is bad.
> 
>>    For an STD or BCP that contains one RFC:
>>
>>       [STDXXX]  Last name, First initial., Ed. (if applicable)
>>                 "RFC Title", BCP/FYI/STD ##, RFC ####, Date of
>>                 Publication.
>>                 <http://www.rfc-editor.org/info/std####>
>>
>>       Example:
>>
>>       [STD72]   Gellens, R. and J. Klensin, "Message Submission
>>                 for Mail", STD 72, RFC 6409, November 2011.
>>                 <http://www.rfc-editor.org/info/std72>
> 
> So if I want to reference Section 3 of RFC 6409, through how many
> indirections will the reader have to go to actual get to the referenced
> text???
> 
>>    For an STD or BCP that contains two or more RFCs:
>>
>>
>>       [STDXXX]  Last name, First initial., Ed. (if applicable)
>>                 "RFC Title", BCP/FYI/STD ##, RFC ####, Date of
>>                 Publication.
>>
>>                 Last name, First initial., Ed. (if applicable)
>>                 and First initial. Last name, Ed. (if applicable)
>>                 "RFC Title", BCP/FYI/STD ##, RFC ####, Date of
>>                 Publication.
>>
>>                 <http://www.rfc-editor.org/info/std###>
>>
>>      Example:
>>
>>       [STD13]    Mockapetris, P., "Domain names - concepts and
>>                  facilities", STD 13, RFC 1034, November 1987.
>>
>>                  Mockapetris, P., "Domain names - implementation and
>>                  specification", STD 13, RFC 1035, November 1987.
>>
>>                  <http://www.rfc-editor.org/info/std13>
> 
> I find it very confusing that the STD itself doesn't have a title.
> 
> It's also confusing that the individual entries do not follow the
> proposed RFC format.
> 
> It's also not clear why we are repeating the STD13 information here.
> 
> So I believe this would better be:
> 
> 
>       [STD13]    "Domain Names", STD 13,
>                   <http://www.rfc-editor.org/info/std13>. Consisting
>                  of:
> 
>                  Mockapetris, P., "Domain names - concepts and
>                  facilities", RFC 1034, November 1987.
>                  <http://www.rfc-editor.org/info/rfc1034>
> 
>                  Mockapetris, P., "Domain names - implementation and
>                  specification", RFC 1035, November 1987.
>                  <http://www.rfc-editor.org/info/rfc1035>
> 
> a) Yes, we'd need names for the STDs and BCPs.
> 
> b) It still makes it hard to reference something specific in the series
> contents.
> 
> This absolutely needs much more work, please don't rush it out.
> 
> Best regards, Julian
> _______________________________________________
> rfc-interest mailing list
> rfc-interest at rfc-editor.org
> https://www.rfc-editor.org/mailman/listinfo/rfc-interest
> 


More information about the rfc-interest mailing list