Re: [tag] Groups - TA Anatomy V0.4 (AnatomyTA-v04.doc) uploaded

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

On 22/09/2007, stephen.green@systml.co.uk <stephen.green@systml.co.uk> wrote:
> I think we are saying the test result should be discarded
> as 'inapplicable', 'fail'(??), 'error' or 'inconclusive'
> if the post condition isn't true or pass.

If you are looking at post test conditions, then the test is complete
and the result has already been determined.

The post test conditions represent the condition of the UUT after
testing is complete. Again a boundary definition issue?
If some part of the UUT could be in more than 1 state for whatever
reason, that is something that can be stated, but is only of interest
to later tests which require that state to be determinate.



>
> We do have a legacy of uncertainty about post conditions
> (I find it unclear how a condition is post and not pre)
> but it is part of things we should align to, I guess.

By definition, for test N, the state of the UUT represents
and may be documented by, that test. No other, its the state
of the UUT after *this* test.
Coincidentally, it is also the pre-condition (prior to test entry)
for the test which is selected to be run next. It may need
to do some UUT conditioning prior to running.


>
> Some just use it as an extension of 'result', others as
> a way to describe preconditions for a subsequent step in the
> flow (a precondition before considering the next step perhaps)

which is another good reason for a glossary with hard definitions
of the terms we are using.


>
> We kind of accept that a post condition is like a result

I don't. See my earlier rationale.

regards



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