Issue 7 - possible resolution

[Francisco Curbera <curbera@us.ibm.com>]

As agreed at the last f2f, my proposed resolution of issue 7 is described
below (#1.) Based on that discussion, we can consider a second flavors of
the proposal, described in #2. I suggest we adopt one of them (#1 is my
preference) and close this issue.

1. Introduce a single <bpel:import> element used to import  XSD, WSDL
definitions. Zero or more <bpel:import> elements may appear as first
children of the <bpel:process> element. Three required attributes must be
included on each <bpel:import> element:

- targetNamespace being imported
- location: where to find the imported information
- importType: identifies the type of document being imported by providing
the URI of the encoding language: either "http://www.w3.org/2001/XMLSchema
", or "http://schemas.xmlsoap.org/wsdl/";.

<bpel:import targetNamespace="..."  location="... document location ..."
importType="XYZ-namespace-uri/>

2. The alternate version of the eliminates the importType attribute.

What is the difference? In option 2, a BPEL processor finds and import and
retrieves the target document without knowing what kind of document that
is. The processor must open and inspect the document in order to figure out
how to process it. On discovering the type of the root element inside the
document, the processor will be able to dispatch a Schema or WSDL processor
to consume the document (which will in turn read the document and now
actually process it.) In option 1, the BPEL processor has not need to read
in the target document, it can just handle the job to the appropriate
(Schema or WSDL) processor.  Otherwise proposals are equivalent, of course.