- From: David Marston/Cambridge/IBM <david_marston@us.ibm.com>
- Date: Thu, 18 Apr 2002 13:28:08 -0400
- To: www-qa-wg@w3.org
On 12 April, Mark Skall proposed:
Test Assertion
A set of premises that are known to be true by definition in the spec.
...
Test Requirement
Same as test assertion
On 4/15, responding to Lofton, he showed a structural approach, with
an example that could be viewed as:
<assertion>
<purpose/>
<scenario>
<required-material>...</required-material>
<stimulus param="value"... />
</scenario>
<correct-response param="value"... />
</assertion>
On 4/17, Sandra Martinez showed an example that looks more like the
(computer-style) grammars we deal with:
"A document consisting of prolog followed by element then
miscelaneous [sic] items is a well-formed document."
Furthermore, Dimitris is apparently going to present another idea
for machine-processable expression of specifications, though I don't
know if it will reach the granularity of the test assertions.
Taking all the above together, it seems that there would be utility
in separating the definitions of "test assertion" and "test
requirement" or inserting a new term. One term would be reserved in
our usage for machine-processable expression of a testable
requirement. A different term would refer more broadly to any
expression of the requirement.
.................David Marston
Received on Thursday, 18 April 2002 13:33:20 UTC