Re: [wsbpel] Gotos Considered Harmful?

[Ron Ten-Hove <Ronald.Ten-Hove@Sun.COM>]

    I believe the assertion being made is that an analyst will create an 
incomplete BPEL model, but nonetheless a model that obeys BPEL's 
structural rules (i.e., no cycles!). The "translation" done by the 
technical modeller is simply a matter of providing the missing detail, 
needed to create an executable process model. (Perhaps converting prose 
into XPath expressions, that sort of thing.)

    It would be interesting to investigate what would be required to 
translate high-level process models to BPEL (which the technical 
modeller would need to complete, of course.) Do we want to take it as a 
requirement to provide such a mapping, to ensure that such a 
transformation is at all possible. (I'm thinking of a one-way, 
mechanical translation. I don't think anything else is really practical, 
but I'd love to be proven wrong!)

    I for one haven't heard of any current process modelling tools that 
can be translated to/from BPEL. Of course, there are a lot more 
opportunities to link process design/analysis tools to BPMSs besides 
simple translation, and it would be helpful if they could be 
standardized (perhaps as extensions to BPEL?). A BPMS could provide some 
very useful services to generate feedback to a modelling tool...

Cheers,
-Ron

Carpenter, Robert E wrote:

>I think that the discussion of human translation of the business analyst's model into a BPEL model via the skeleton makes a lot of sense, but my questions are: 
>
>(1) Does anyone have an example of a working translation process for other "higher order" BPR tool to/from BPEL ? An example would be SCOR level 3[or possibly SCOR level 4] <---->BPEL.
>
>(2) Does anyone know the "state of the art" around automated conversion between these two modeling paradigms - i.e. SCOR <---->BPEL ?
>
>
>Rob Carpenter
>Intel Corporation
>
>  
>