"Williamson, Rusty" wrote:
> [snip]
> However, do whatever works for you, yes? If you're using RUP however, you
> almost have to go with the OMG Spec or Position A [snip]
> Chop up use cases into incomplete bits and they
> fall apart in RUP. They won't support iteration planning well, or realize
> the test model, or work for estimating man-hours (see Schneider), or do well
> in the use-case realization to analysis classes transition.
[snip]
Mr. Williamson, I'm not sure that we're talking about the same thing.
The issue is not about making smaller use cases -- the issue is about
how use cases are covered by scenarios.
Given that I am not suggesting partitioning use cases into small,
meaningless pieces of work, do problems with the RUP still arise?
Going from Quatrani (VISUAL MODELING), a use case flow of events (FOE)
starts with some stimulus to the system, causing some pattern of
responses, followed by selection of an appropriate subflow, followed by
the end of the use case. Exceptions may be thrown during the sequence
of events. Details of subflows and exceptions are documented with the
FOE. I accept this model for purposes of this discussion.
The issue is: must scenarios always show a complete path through the
FOE? Position A says yes. Position B says no.
It doesn't take a complicated use case to require several quite long
scenarios for "happy day" coverage of the FOE. Then too, there is no
obvious reason why there shouldn't be branching structures in the
subflows, further multiplying the number of unique paths through the
FOE. Each exception also multiplies the potential number of unique
paths, and may also contain branches. You would be forced to use
permutations to write an expression for the number of paths.
Scenarios threaten to become both long and numerous if every scenario is
a complete path. Sequence diagrams can grow so long and wide that they
cease to be printable, and the mind can't cope with them. As a
practical matter, I don't see how you can do enough scenarios for good
coverage of the FOE.
The very existence of exceptions that may "abend" the FOE also fuzzies
up the notion of a "complete" path.
Thus, my confession: I'm a B guy.
Scenario coverage is still nontrivial for B guys, but it becomes more
manageable, I think. This could be the source of another good
controversy.
That'll do for now. I've gotta get up and teach more UML in the
morning. Thanks to everyone for the discussion -- it's been amazing!
-Eric
************************************************************************
* Rose Forum is a public venue for ideas and discussions.
* For technical support, visit http://www.rational.com/support
*
* Admin.Subscription Requests: [EMAIL PROTECTED]
* Archive of messages:
http://www.rational.com/products/rose/usergroups/rose_forum.jtmpl
* Other Requests: [EMAIL PROTECTED]
*
* To unsubscribe from the list, please send email
*
* To: [EMAIL PROTECTED]
* Subject:<BLANK>
* Body: unsubscribe rose_forum
*
*************************************************************************
