RE: Prescription levels

[david_marston@us.ibm.com]

>its a slippery slope
to expect the TA to tell "what you can conclude about the product".
 
Agreed. But I think we need to see some
pilot projects using this TA structure to determine the role of the Prescription
Level.

>So far the TC has been
careful staying away from associating TAs with any conformance conclusions.
These are left to test cases / test suites (resulting in "pass"
or "fail"). 
>All we have with a TA,
is an indicator of fulfillment / violation of some spec requirement.

So you're already on that slippery slope.
I wouldn't argue for the abolition of the Prescription Level, and if it's
there, it has the effect of spelling out the consequences of the predicate
being true. (Also remember that a TA can be used to draw a conclusion about
a measured property of the product, which is a Prescription Level of either
MAY or Information Only. In your examples, a TA that asserts the property
of being medium-sized for a widget is the type I mean. There is no conformance
conclusion about the fact of being medium-sized, but there is a conclusion,
which in turn can be used in the prerequisites of other TAs.)

>Back to our concern,
I take it from your email that.. you agree with previous thread consensus....

Yes.