Re: [6tisch] removing the 'e'

[Tom Phinney <[e-mail]>]

Just to be clear about the ISO/IEC approach, there are two places where 
a dated reference might be specified. The first is simply within the 
Normative References clause of a standard, where the specific edition to 
which the standard refers can be cited when that is deemed appropriate. 
For IETF documents, where a different document number is assigned for 
each revised republication, the document number alone suffices. However, 
for standards from other standards development organizations (SDOs), the 
specific year of the edition usually is required when it is deemed 
necessary to uniquely identify a specific edition of the cited standard.

The second place where a dated reference must be specified is when the 
text of the standard cites a specific subclause of the referenced other 
standard as normative text. In that case it is necessary to cite the 
year of the referenced standard within the reference so that an 
oversight during future editing of the citing standard does not 
inadvertently reference the wrong clause of a revision of the referenced 
standard, which could otherwise occur when the base edition of the 
referenced standard is updated within the Normative References clause of 
the citing standard.

A hypothetical example from the current documents shows the various 
difficulties:

- Current normative references within 6tich documents would be to IEEE 
802.15.4:2011 or to IEEE 802.15.4e:2012. Both would be listed in the 
document's Normative References clause.

- A textual reference could take the generic form "IEEE 802.15.4", which 
would be appropriate when referring to aspects that are common among all 
editions and revisions to IEEE 802.15.4.

- A second form of textual reference would be to either "IEEE 
802.15.4:2011" or "IEEE 802.15.4e:2012", which designates the specific 
referenced source document but not a specific subclause within that 
cited document.

- The third form of textual reference is to a specific subclause of a 
specific document. In that case the reference form must include the 
date, such as "IEEE 802.15.4e:2012, 4.7.2" to reference subclause 4.7.2 
of IEEE 802.15.4e:2012. Including the date at the point of reference 
ensures detection of future editing errors during maintenance of the 
citing standard.

Note that the specific syntax of an undated or dated reference depends 
on the SDO that is writing the standard. The above shows the required 
syntax for ISO/IEC standards, in which the document number itself often 
contains a hyphen ("-"), thus making the hyphen unusable as a date 
separator. IETF and IEEE each have somewhat different requirements. The 
rules about inclusion of spaces within the citation also differ by SDO.

Regards,
-Tom
=====
On 2015.03.31 05:21, Brian Haberman wrote:
> Pascal,
>
> On 3/31/15 2:58 AM, Pascal Thubert (pthubert) wrote:
>> Thanks a bunch Tom!
>>
>> I think that the case must be brought to the attention of the ADs,
>> CC'ing Brian. I'm unclear what the IETF best practice is for external
>> documents such as IEEE, but the IEC approach seems very well though
>> through. If we do not have such a detailed BCP, then this should be
>> exposed to the IAB.
> I am not aware of a BCP that discusses references to external documents.
>   Format-wise, that is the purview of the RFC Editor.  I am not sure why
> the IAB would care about references...
>
> I am generally leery of a blanket reference to the generic
> specification.  The WG has developed its solution based on a particular
> version of the IEEE specification, so I would expect to see references
> annotated by the year of the 15.4 spec.
>
> Regards,
> Brian
>