Re: what's normative [was: Re: processing plan for SpecGL LC issues]

At 05:44 PM 4/13/2003 -0400, you wrote:
>[...]
>Food for thought.....
>
>>>The Checkpoint Priorities:  Are these Normative?  (#106)  (May be moot, 
>>>depending on resolution of definition of normative)
>>
>>Question.  Does he mean the definitions in 1.7, or the priorities 
>>themselves that are associated with each checkpoint.
>
>Good question.  I read it as section 1.7.  But, it really doesn't matter, 
>since we should address both cases.  For section 1.7,  it is not normative 
>(if we stick with last week's agreement).

I agree.  One can conform to SpecGL (or NOT!) at A, AA, AAA, without having 
a clue what Priority 1 or 2 or 3 actually mean (which is what 1.7 is about).

>As for the priorities themselves as associated with each checkpoint - I 
>think that they need to be normative, since they have conformance 
>consequences, in that they identify what belongs in A, AA, AAA.

I tend to agree, but am not completely sure yet.  On the one hand, they 
don't really prescribe any requirements on a subject specification.  On the 
other, they do prescribe which conformance category a subject specification 
will fall into as it passes/fails checkpoints.

>Thus, the Checkpoint is normative as is the Conformance Requirement 
>statement(s).

Not so sure about the checkpoint statement.  We have (in the past) 
described it as a "title" for the checkpoint's real stuff -- the 
Conformance Requirements.  One could argue either way, whether the title of 
a normative bit is part of the normative bit, or not (oops ... I meant 
NOT!).  If we decide that the Priority label is normative, then I have no 
problem (see no inconsistency) if the checkpoint statement isn't.

-Lofton.

Received on Sunday, 13 April 2003 23:28:45 UTC