[rfc-i] rfc-interest Digest, Vol 196, Issue 22

[John C Klensin]

--On Tuesday, February 23, 2021 22:14 -0800 Christian Huitema
<huitema at huitema.net> wrote:

> On 2/23/2021 9:15 AM, John C Klensin wrote:
> 
>> (2) References from RFCs.  These need to be, at least by
>> default, exact.  The flexibilities that are an advantage with
>> the above can be a liability here.  Why? Because it is not
>> unusual for the substantive content of an I-D to change as
>> work evolves and consensus emerges (or doesn't).  If, for
>> example, version NN of an I-D said "the outside of the
>> bikeshed MUST be painted blue", version NN+1 said "the
>> outside of the bikeshed MUST be painted lime green" and
>> explained why, it would be very important if an RFC that
>> referenced it for color choices point to the version the
>> RFC's author intended and not some other version.  Would it
>> be important for the reader to find out that there is a later
>> version in which things might have changed? Sometimes, but
>> only the document author is likely to know.
> 
> Or, take the example of an Internet draft that ends up
> abandoned. References to such drafts are not rare. Now, what
> is also not rare is for the author of the abandoned draft to
> come back and "tombstone" it. You certainly don't want the
> reference to magically be updated and point to the tombstone.

Indeed.
    john