Re: Test case metadata...

[david_marston@us.ibm.com]

Hyunbo Cho asks:
>By the way, can I ask a question about "test case metadata"?
>Does it mean by "test requirements"?

No. Test case metadata is a set of information items provided for
each individual test case. If the tests are organized into groups,
there could also be per-group metadata. The main uses are:
1. To allow setup of automated testing on several platforms
2. To allow (permissible) filtering of the test cases to just the
   set needed for a particular test subject and test run
3. To provide information that will appear in various kinds of
   reports, both about the suite and about particular test runs
4. To support a platform-neutral notation of scenarios for setup,
   execution, result evaluation, and cleanup
5. To support modularity of test harnesses
If the metadata is in XML, then XSLT can be used to transform it
into platform-specific execution scripts and other useful
material.

Test case metadata could be written before the test cases, which
would make it a very detailed, case-by-case statement of the
"requirements" for each test. When I hear "test requirements", I
think of something more global, like trying to describe what forms
of testing are needed and objectives of the test suite or whole
testing regime. For example, if measuring performance is an
objective, that drives some requirements for capturing times and
the like. Do others define "test requirements" differently, and
specifically in a way that involves test assertions?
.................David Marston