--On Sunday, October 23, 2022 17:44 +0200 Carsten Bormann
<cabo@xxxxxxx> wrote:
> On 2022-10-23, at 17:32, John C Klensin <john-ietf@xxxxxxx>
> wrote:
>>
>> So, to repeat my
>> point, Miles's question about "how we got away from writing
>> compliance tests..." appears to be based on an incorrect
>> premise because we never did write such tests.
>
> Well, "we" did get away from compliance tests in the
> 1980s/1990s by switching to Internet technologies, which were
> emphasizing interoperability over compliance. (And, recently,
> we have learned to significantly improve interoperability in
> some cases by agreeing on test vectors, which are often what
> compliance tests would distill to…)
Certainly, if one holds up OSI and pre-OSI PSTN-based data
technologies as part of what "we" did, then, yes, the shift to
TCP/IP carried with it the abandonment of explicit compliance/
conformance tests are part of the protocol suite and definition
methods. Even then, for many standards, the compliance tests
were designed and administered by supposedly neutral third
parties and not by the standards bodies -- just as the long
tradition of organization interoperability testing for the
Internet and its applications were part of separate
organizational structures and not part of the IETF (and
predecessors) standards process (although we certainly learned a
lot from each other).
On the other hand, for those of us who were fortunate enough to
have come down the ARPANET-> Internet path, or even the
BITNET/EARN/NETNORTH, CSNET, etc., ones -- seeing OSI and its
ITU and ISO predecessors as something to study and learn from
but not to adopt or use in significant or direct ways -- a
different response to your question (with apologies to those who
recall the end of the sentence and interpret it as reverse
racism), "what you mean 'we'...?"
best,
john
