tag message
[Date Prev]
| [Thread Prev]
| [Thread Next]
| [Date Next]
--
[Date Index]
| [Thread Index]
| [List Home]
Subject: Comments on previous comments on (1) IUT + TA, (2) pre-requisites
- From: "Durand, Jacques R." <JDurand@us.fujitsu.com>
- To: "TAG TC" <tag@lists.oasis-open.org>
- Date: Tue, 6 Nov 2007 18:52:38 -0800
Some comments on two points:
1. IUT and TA:
"A best practice technique is to list the
IUTs to which a TA applies using words
like 'applies to ...'
Consensus that IUT should not be required as part
of the individual TA"
I believe there are here different concerns, that
may need to be addressed separately in the guide:
(1) IUT vs COnformance Target
The two should probably be presented as
orthogonal - generally not of the same level of granularity,
but could overlap.
In WS-I Profiles, you have for example TAs that
apply to a [Web service] INSTANCEs
which clearly match to real world products, that
could definitely also be the target of
conformance clauses. But you also have TAs that are
targeting much finer granularity entities
such as [message] ENVELOPE.
(2) Not having to mention the IUT for many
TAs.
I believe that an IUT is an essential part of a TA
(otherwise WHAT is the testing about ?)
BUT the IUT may not need be mentioned in
every single TA, assuming that it can be
made somehow "global" to many. As long as the user
has a means to figure out what the IUT is
(what class of objects) for a TA, then we are
fine.
(3) A same spec req may apply to different
IUTs.
Right - but I feel that a TA cannot afford the same
fuzziness.
For example, if a spec requirement says "error
messages must contain a reference to the faulty
message", then this can be interpreted either as a
requirement for either (a) the error message
artifact itself, or (b) the message handler that
produces such messages.
It think it is OK to write two TAs here, one
for each IUT, or just one, depending which IUT is
of interest.
2. Sequences of Test Assertions, and
pre-requisites
"Make it the rule to keep each test assertion
singular and atomic but recognize that
there are exceptions
when an assertion depends on another."
Rule: keep assertions singular and
atomic.
Yes, singular and atomic in that the
expression that evaluates the test outcome does not
computationally depend on other TA outcome (i.e. can be evaluated without having to resort to the test outcome
expression of another TA)
I think that was the prime concern from Patrick and
Sun.
Concretely, if TA1 is a pre-requisite for TA2, this
means that the test case that exercises
TA2 can still be executed independently from test case for TA1, but that the outcome of TA2 may not be meaningful -
i.e. the logic of TA2 testing
was assuming that TA1 was successful for the same
IUT.
E.g WS-I TAyyyy on presence of some elements in
SOAP header, has pre-req TAxxxx on SOAP
extension compliance of the header.
If TAyyyy fails for IUT 123 you can infer
that 123 was not conforming/adhering
to the requirement addressed by TAyyyy
ONLY IF TAxxxx was successful for this
IUT.
-Jacques
[Date Prev]
| [Thread Prev]
| [Thread Next]
| [Date Next]
--
[Date Index]
| [Thread Index]
| [List Home]