Stream: conformance
Topic: expectation when no extension.
Eric Haas (Apr 26 2018 at 21:14):
What is the expectation if no expectation-extension is used in CapStatement when defining the capabilities of a desired system specification such as for an IG?
Lloyd McKenzie (Apr 27 2018 at 00:13):
I presume it's "shall"
Eric Haas (Apr 27 2018 at 00:22):
Presume is like assume :-) . Suspect we need a GForge for this
Lloyd McKenzie (Apr 27 2018 at 00:39):
Though if it's a "SHALL" by default, I guess that makes expectation-extension a modifier?
Eric Haas (Apr 27 2018 at 00:40):
I would rather keep ismod out of it. I think it should be undefined.
Eric Haas (Apr 27 2018 at 00:40):
Which means you need to use the extension if you want to define it/
Lloyd McKenzie (Apr 27 2018 at 00:41):
Well, when a system declares it, it's pretty much an "I do support" - or the declaration is useless for clients
Lloyd McKenzie (Apr 27 2018 at 00:41):
An alternative is we promote it to core
Eric Haas (Apr 27 2018 at 00:41):
I watched the other WGs struggle with IsMod btw nobody knows how to apply it
Lloyd McKenzie (Apr 27 2018 at 00:43):
Does inclusion of the element with any of its allowed values cause the interpretation of the resource ignoring the resource to not be true
Eric Haas (Apr 27 2018 at 00:43):
We are not talking about what system support here, we are talking about expectations
Eric Haas (Apr 27 2018 at 00:45):
you need to define all the expectations
Lloyd McKenzie (Apr 27 2018 at 00:45):
The resource is used for both purposes. You're proposing that it's precise when defining system capabilities, but fuzzy when defining desired capabilities?
Eric Haas (Apr 27 2018 at 00:46):
I think the extension only makes sense for defining expectations.
Eric Haas (Apr 27 2018 at 00:48):
So Yes.
You are proposing everything is a Shall and mark the exceptions.
Eric Haas (Apr 27 2018 at 00:50):
Which means the extensions should be "reducedExpectations" or rather two extensions May and Should
Eric Haas (Apr 27 2018 at 00:52):
I really don't care which approach ( except for the ismod part which is not needed for an expectation) but need to be clear and allow the extension everywhere
Eric Haas (Apr 27 2018 at 00:54):
promoting it to core is unwise since you would need an element for each element - more economical to keep as an extension forever.
Lloyd McKenzie (Apr 27 2018 at 05:33):
But the point is that those extensions change the interpretation from what would be true if they were ignored - that makes them modifier extensions
Eric Haas (May 01 2018 at 15:15):
And definately no on IsMods for something that is an expectation for behavior - its is defining proposed behavior an not actual behavior.
Last updated: Apr 12 2022 at 19:14 UTC
