about issue i031

["Durand, Jacques R." <JDurand@us.fujitsu.com>]

I withdraw my Proposal #3 (composite TAs).

 

Reasons:

- the notion of composite TA seems to go beyond the scope we assigned for TAs: "...singular, atomic statements more limited in scope and purpose".

- We know there is a need for grouping TAs in more meaningful "test entities" e.g. addressing a conformance level or profile, or a complex property of an implementation ("mobile widget", "medium-size widget"...) but that is beyond the scope of this guideline to define these test entities (at best, to enable them using tags, headers...).

 

For the same reason I am not convinced we need to define "global predicates" (Stephen proposal #4).

 

On the other hand, we know we need to address complex prerequisites, like: TA1 is applicable on a target "t" if TA2 evaluates to true for "t" and if either TA3 or TA4 evaluates to true for "t".

So we might need to introduce a notation for this, something like:

 

TA id: TA1

target: t

prerequisite: TA2(t)=true AND ( TA3(t)=true OR TA4(t)=true )

 

(NOTE: TA2(t)=true is not exactly same as TA2.predicate(t)=true, because if the prereq of TA2 is false, then TA2(t)=notApplicable while TA2.predicate(t) might still be true)

 

But complex prerequisites are a different problem: the TA is still addressing an atomic normative statement. Only its applicability is complex and involves other TAs. Likely this notation can be used beyond prerequisites but that is also another issue.

 

Therfore I am backing the more modest Proposal #2 for i031.

 

Jacques