On 6/20/24 03:35, Mario Loffredo wrote:
Hi Andy,
[snip]
On 6/18/24 04:29, Mario Loffredo wrote:
Hi Andy,
what does it mean "interoperable implementation" in RegExt context ?
Best,
Mario
It means the protocol/extension works as expected. If Bob were to
create a client with extension Foo and Alice were to create a server
with extension Foo, but the expected features of extension Foo did
not work between then they do not interoperate. Of course, "did not
work between them" can be a bit subjective but that is why a written
report is required.... to pull out the ambiguities.
[ML] Is it the same as requiring the "Implementation Status" section
to include at least one client and one server implementation?
We could do that, or some combination thereof. RFC 7942 which covers the
implementation status section list the use of the wiki as an
alternative. The benefit of the wiki is that the information doesn't get
lost when the document turns into an RFC. The benefit of the
implementation status section is that it is with draft as it is reviewed
by the various reviews of the IETF. We could opt to do one or the other
or both or a hybrid (link in the implementation status section pointing
to the wiki report). One benefit the wiki seems to have is that the
implementers can update their reports and the information does not need
to go through the draft authors. The wiki is also probably easier to format.
If we go with the implementation status section, we should require the
"coverage" and "implementation experience" parts be more thorough,
including a list of the type of interoperability tests conducted and how
they were performed. If we use the wiki, I think the reports should
include the same information that would go into the implementation
status section.
-andy
_______________________________________________
regext mailing list -- regext@ietf.org
To unsubscribe send an email to regext-le...@ietf.org
