OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

ebxml-bp message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]


Subject: Part 2 of Refactoring action item (implementation enhancement): New treatment for embedding, initial remarks on traversal in monitoring/verification tasks


Hi

In the interest of getting 90% of the refactoring proposal out to the
TC, here is the second part of the refactoring, mainly dealing with
embedding (or nested) BTAs. This replaces the current syntax (
@onInitiation attribute = true for a Transition) with some new
constructs, and introduces a placeholder that is currently used to
indicate how/whether status from subordinate BTAs (or
CollaborationActivities) is to be  conveyed back to the superordinate
BTA. This allows designers to specify that a Failure from a subordinate
process should cause a signal to sent instead of a Response, for
example, and return a Failure status code to the collaboration
environment.

I also was prompted by some remarks by John Yunker to begin to discuss
how the statuses, document language, and xpath language "events" relate
to the traversal of the BPSS states within the monitoring/verification
task. This discussion is incomplete, because it does not treat the
questions of delay/nondelay on transitions very completely. (Time only
enters as an event pertaining to signal generation (and status setting)
that percolates up through timeouts on timeToXXX.) 

Nevertheless, this is worth getting out to get reviewed to see whether
it helps implementer's understand BPSS declarations and how they can be
used in constructing monitors and verifiers of BPs.

I did look at UML 2.0. I have not spent a lot of time discussing
fidelity to UML models for StateCharts and ActivityDiagrams. I have come
to the belief that fidelity to these models has not been a main concern
in earlier versions, nor is fidelity likely to become more important
with 2.0. To some extent, the abstraction level of BPSS does not really
require a full Activity modeling notation, and selective borrowing seems
to be a more useful relation to those models.

Dale Moberg

Refactoring Embedded-Nested States.doc



[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]