Mail Archives: djgpp/2017/10/09/04:36:39
tl;dr RFCs classified as Proposed Standards are indeed standards for
practical purposes, although there is no enforcement agency. An
Internet Standard is just a relabeling of an existing RFC. The status
of the cited RFC 1855, however, is Informational, not Proposed Standard.
The long version includes historical rationale for why it works this
way. Please reply to me if you want to discuss; Reply-To set (but
this list may reset it; please check the addressees if you do reply).
Mateusz Viste writes:
> 1. RFC are no standards, otherwise they wouldn't be called "RFCs"
> in the first place
That's not a realistic description of actual practice.
RFCs are technically classified as "Proposed Standards", "Best Current
Practice", and "Informational". "Internet Standards" are simply
redesignations of particular Proposed Standards as Internet Standards,
without changing any text. Neither Proposed Standards nor actual
Internet Standards are enforced, except by the acceptance of the
community and refusal to communicate with nonconforming applications.
In that sense you might consider both to be "non-standards".
However, as formal descriptions of communication processes that most
of us strive to conform to, Proposed Standards are standards indeed.
It is acceptable to criticize an implementation that claims
conformance to a Proposed Standard on the grounds that it does not
implement all of the REQUIRED behaviors, as well as all of the
RECOMMENDED behaviors where a good reason for the lack is not
presented. It is acceptable, though often impractical (think Yahoo!
mail or Gmail, or Reply-To- munging mailing lists like this one ;-),
to refuse to communicate with a badly implemented or non-conforming
application.
The original intent was that Proposed Standards might be superseded by
a better proposal of a similar protocol, which would be elevated to
Internet Standard instead of the original. In practice it turns out
that the "working code and rough consensus" process produces products
robust enough, and takes long enough, that Proposed Standards attract
near-universal conformance in most cases, and rarely are actually
elevated to Internet Standard.
> 2. The very same RFC you mention does say:
>
> "this Guide offers a minimum set of behaviors which organizations and
> individuals may take and adapt for their own use."
> ^^^
Indeed, RFC 1855 is also explicitly classified as "INFORMATIONAL" in
the heading, which means the same thing.
For completeness, "Best Current Practice" RFCs describe emerging
consensus on
- implementation techniques,
- on local handling of data received or to be sent (Proposed
Standards define behaviors needed for inter-system
interoperability, and thus should not discuss purely local
behavior), and
- about the best configuration of optional features (and sometimes
changed opinion about whether to implement recommended features by
default).
Hope this helps!
Steve
- Raw text -