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

 


Help: OASIS Mailing Lists Help | MarkMail Help

wsbpel message

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


Subject: Re: [wsbpel] Issue - 166 - Proposal To Vote


Andrew,

    The clause Yaron added ("MUST be executed as if, for the duration of its execution, it was the only activity in the process being executed.")  provides isolation; atomicity is already required. These two properties result in consistency (a weak form, admittedly, but we haven't any other form of consistency defined for BPEL process instances). (It might be worth visiting these issues with respect to, say, transition condition evalution.)

    Durability hasn't been mandated by any other part of the spec, and I doubt we could arrive at an adequate defintion without getting dangerously close to mandating implementation details fair outside our domain as creators of a portable business process execution vocabulary.

    A simple "factoid:" I have a workflow management product that provides for full recovery of process state in the event of engine shutdown or crash (more easily said that done!)[1]. I also have customers who turn that feature off, because they want the extra speed that buys them, and, in their applications, it makes no sense to recover from such a crash. I hope we would not restrict BPEL engine  implementations such that I will have some grumpy customers, and no wiggle room in an overly prescriptive spec.

-Ron

[1] Forté Conductor Workflow Engine, now known as the Sun Integration Server, EAI Edition, Process Engine.

andrew.francis@mail.mcgill.ca wrote:
Hello Danny:

  
"The assign activity MUST be atomic, that is either all
<copy> elements  are performed, or none are.
    

I have two suggestions. The first is to state that
assign exhibits ACID properties. This is the natural
language narrative. Secondly, write an appendix, which
in a more formal language, describes the behaviour of
assign. This way we get rid of the ambiguity for the
BPEL engine designer.

Cheers,
Andrew



To unsubscribe from this mailing list (and be removed from the roster of the OASIS TC), go to http://www.oasis-open.org/apps/org/workgroup/wsbpel/members/leave_workgroup.php.

  

S/MIME Cryptographic Signature



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