Closed versus open lists of test assertion

[stephen.green@systml.co.uk]

My findings related to use of RDFS/OWL for test assertions
are mentioned here:

http://lists.oasis-open.org/archives/tag/200801/msg00038.html

The main point is that if using OWL as an ontology language
for defining test assertions the problem arises that OWL has
an open list philosophy which seems inherently unsuitable for
test assertions.

By 'open list'/'closed list' I mean this:
In a closed list there is certainty (or at least confidence)
that the list at some point is complete. A by-product of this
is that a closed list tends to have the means to remove an
item from the list because the complete content of the list
is somehow defined (say with an index). There is a change
procedure for adding and removing items from the list so at
any one time the list contents can be known and is usually
deliberate.

With an open list there is no limit to what might be in the
list. There is no list anywhere of all list items. And with
OWL as an example of this, there is no clear way to remove
an item from the list as a result.

I take it that most scenarios of test assertions require a
closed list as described above: to be clear what is and
what is not a test assertion, even though this list can be
changed or extended through a change process. This seems
to not only limit use of OWL but for the matter in hand
(grouping) seems to need an explicit understanding and to
require limitations on what is appropriate for any listing
mechanism used for test assertions. It also suggests that
to keep the outermost set of test assertions for a set of
targets well defined ('closed') there always has to be
some outermost grouping construct. Perhaps this is likely
to be the conformance clause(s) but then there is a timing
issue to consider - that the conformance clauses might not
have been written when the test assertions are written or
that the conformance clauses might have been written before
the test assertions and so be unaware of them.




-- 
Stephen D. Green

Partner
SystML, http://www.systml.co.uk
Tel: +44 (0) 117 9541606

http://www.biblegateway.com/passage/?search=matthew+22:37 .. and voice