Comment on changes in the TA model and markup

[Stephen D Green <stephengreenubl@gmail.com>]

I've been looking at the changes the TC is proposing for the
Test Assertion model and markup and on the whole I welcome
the simplification of the test assertion 'Set' object. It seems a
good principle to me to reflect back on the model and markup
once it is agreed and see whether all of it or only part of it is
likely to be widely implemented and prune out what is not
likely to be implemented but merely adds model/schema bloat.

I do question whether the scenario supported hitherto - the
reduction of redundancy in applying a prerequisite to a whole
set of test assertions so that these can all be ignored as a
block when it is known that that prerequisite will not be met or
is perhaps inapplicable - whether the removal of the common
prerequisite from the TA Set reduces the support for this case.
On the whole I could still support this use case by making a
single TA's predicate refer to a TA Set and adding the prereq
to that TA. This is possible as long as a TA Set has an ID and
that particular implementations or binding profiles allow this
ID to form part of a TA predicate expression. So I'm still fine
with this.

I took it that the only changes to the markup are as found in
the attached schema. DES had previously applied this schema
to represent an OASIS UBL example calculation model and I
believe the resulting Test Assertion Markup Language instance
needs only the very minor change of the removal of the TA
document header element while keeping its 'common' element
child as a child directly descending from the simplifiied TA Set.
I attach the same which if the changes to the markup stay now
as they are until they are part of a candidate OASIS standard
I would then like to resubmit to the UBL TC as a comment.

Best regards

Steve

----
Stephen D Green
Associate Director
Document Engineering Services

ubl-ta-draft-1-1.zip