Satish Thatte wrote:
>I don't buy the three days vs one day example - I don't see which
>interfering shared variable access is essential in that example.
>
Sorry for the misunderstanding; I wasn't referring to the example in the
spec; I was referring to the use of the travel agency example across
many such process specs. What I meant (but did not spell out explicitly)
was that suitable patterns of shared variable access could result in
serialized execution of concurrent scopes. I don't believe static
analysis of shared variable use can accurately detect this problem,
though it may be able to warn about potential problems.
>What did you have in mind for the handling of concurrency faults?
>Retry? Failure? Some custom code?
>
This is really a specific case of the more general problem of fault
handling / recovery in BPEL. The language should have decent support for
this. In this case, I might follow a typical pattern: compensate for the
activities completed in the failed scope, then retry it. Nothing fancy
like custom code, as much as I admire Edwin's <exec> tags :-).
Cheers,
-Ron
