Re: [tag] some thoughts and comments from reading a month worth ofmessages

["Dave Pawson" <dave.pawson@gmail.com>]

On 22/09/2007, Durand, Jacques R. <JDurand@us.fujitsu.com> wrote:

> Consider the specification requirement:
>
> " When API function F(x) is called with argument value greater than MAX,
> then the result SHOULD be same as F(MAX)".
>
> Just using the above statement as "TA prose", will leave TA users
> unclear as what are the pass and fail conditions for this TA: the use of
> keyword "SHOULD" here might lead to conclude that  (x > MAX) AND (F(x)
> != F(MAX)) => pass, since this is authorized (though not recommended) by
> the specification.

This is a variant on the clear pass fail result. The test may have failed
but the outcome, as agreed with the spec author, could be simply a warning,
since the requirement is a SHOULD.

Not addressed by this group, what constitutes an overall result.
Simplest, a clean 'and' of all results.
Variants, as illustrated. Pass, with n warnings.
The spec author / QA need to agree what constitutes an overall
pass. suggest address it, but definition is out of scope for this group.


>
> But the rationale behind writing a TA for this statement, is that it is
> supposed to only verify if the recommended behavior/feature is supported
> by the IUT or not. In such case, (x > MAX) AND (F(x) != F(MAX)) => fail,
> while: (x > MAX) AND (F(x) = F(MAX)) => pass.
>
> How to convey this test logic in a "prose" TA?

"If the test fails, a warning shall be issued, reporting the value"




> NOTE: when there are many spec requirements (either MUST or SHOULD) of
> the same style for a large number of API functions, they could all share
> the same additional meta statement (b), so as to avoid repetition.

Or each could be treated individually. Down to the spec author who knows
why they are SHOULD and not MUST.


-- 
Dave Pawson
XSLT XSL-FO FAQ.
http://www.dpawson.co.uk