On 5/21/24 04:20, Rob Wilton (rwilton) wrote:
Broadly speaking I’m in favour of IETF focussing its energies on work that will get deployed.
That sounds good at first, but I think caution is appropriate
here. Ideally, IETF has *influence* over what gets deployed; its
recommendations are valued enough by implementers and vendors that
they consider building or adapting products to IETF
specifications. I do think this is true to some extent, but IETF
can also miss - it sometimes does develop specifications that <
2 parties want to implement.
The other caveat is that IETF should consider the needs of the
broader Internet community. In particular it shouldn't merely
follow what one or more Big Vendors does, because no Big Vendor
has the entire Internet's interests in mind.
(Sometimes a Big Vendor produces a protocol that almost meets the needs of the broader Internet community, and which IETF can tweak to be more broadly applicable. That can be a happy outcome, though is not necessarily so.)
But none of this has anything to do with interoperability tests. Interoperability tests can provide some minimal assurance that a protocol specification is good enough to permit interoperability of independent implementations, but they provide ZERO assurance of success of the protocol in the "marketplace", or if you prefer, the Real World.
Keith
