Re: Making RFC2119 key language easier to Protocol Readers

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 




--On Tuesday, January 15, 2013 00:19 -0600 Dean Willis
<dean.willis@xxxxxxxxxxxxx> wrote:

> On Jan 5, 2013, at 10:03 AM, John C Klensin
> <john-ietf@xxxxxxx> wrote:
> 
>> And, again, that is further complicated by the observation
>> that IETF Standards are used for procurement and even for
>> litigation about product quality.   We either need to accept
>> that fact and, where necessary, adjust our specification
>> style to match or we run the risk of bodies springing up who
>> will profile our Standards, write procurement-quality
>> conformance statements for their profiles, and become, de
>> facto, the real standards-setter for the marketplace (and
>> obviously do so without any element of IETF consensus).  
> 
> I'm not sure that's not a good thing. Witness for example the
> work SIP Forum has done with the SIPConnect standards, which
> have made it MUCH easier to order a box that will work with a
> SIP Trunking service.

Dean,

Historically, there are many cases where certification
arrangements have worked out well, many cases where they have
worked out badly, and many cases where evaluation of the
certification process has, itself, been controversial.   That
really isn't the point.  The issue is locus of authority in
practice because, if certification is important in the
marketplace, then the real standard is ability to conform to and
pass certification, not conformance to the published Standard or
the ability of implementations built to that published Standards
to interoperate.  

One can imagine a certification test that fatefully follows the
published Standard and verifies every feature and requirement
_and_ that that either the published Standard is well enough
written to not allow alternate interpretations that would not
interoperate or the certification process includes adequate
interoperability tests.  In that case, the difference between
"conform to (i.e., pass) certification test" and "conform to
published Standard" is mostly a matter of aesthetics and
perceptions of control.   

However, IETF has often been very concerned about change control
and, no matter how good everyone's intentions, successful
certification activities often switch the locus of that control
(sometimes, so do unsuccessful ones and ones that are mandated
by some authority).  In addition, some people would suggest that
the near-perfect certification test outlined above bears a
strong resemblance to provably-correct computer programs: the
proofs are possible and work well for very simple cases,
primarily those that are designed an constructed with proof in
mind and not for more complex and organic situations.

    john



[Index of Archives]     [IETF Annoucements]     [IETF]     [IP Storage]     [Yosemite News]     [Linux SCTP]     [Linux Newbies]     [Fedora Users]