RE: [wsbpel-implement] Portability and Interoperability Testing

["Kristofer Agren" <kagren@pakalert.com>]

I agree that this needs to happen in order for BPEL to gain acceptance in
the industry. Here are some of objectives that spring to mind:

1. Provide the tools for implementers to validate their implementations,
this could include:

1.1. Defining a set of BPEL processes/use cases that will collectively cover
the functionality of the entire BPEL specification. These test cases should
be delivered in conjunction with new the release of new BPEL specifications.

1.2. Define the expected results when running the BPEL processes above,
which could be done by either supplying the message contents that are
involved and their flow, or providing actual implementations on a central
site so that the test processes can be run with different input.

1.3. Either providing a set of WSDL definitions for the web services that
will take part, or even better, provide actual implementations at a central
site so that there will be no misinterpretations when implementing the web
services involved in the business process.

2. Provide some ways for, for lack of a better work, certification. This of
course implies that some independent entity is available to do the
certification. While certification is a huge effort by itself, without it,
there will be no way to tell if a given implementation is in fact compliant.

3. Interoperability. This would imply cooperation on behalf of all, or a
large majority of, BPEL implementers. How do we get implementers to ensure
portability amongst themselves?

Regards,

Kristofer

-----Original Message-----
From: Ron Ten-Hove [mailto:Ronald.Ten-Hove@Sun.COM] 
Sent: Friday, February 13, 2004 1:38 PM
To: bpel implementation
Subject: [wsbpel-implement] Portability and Interoperability Testing

BPEL implementors:

    I've been giving some thought lately to the issues of testing or 
demonstrating compliance to the WS-BPEL specification. As we have 
discussed in the past this really falls into two categories: portability 
and interoperability.

    Historically there have been several approaches to testing such 
compliance. Our situation is complicated by dependencies on 
implementation-specific messaging infrastructure, so we have some other 
issues to deal with before we get to the main event. What approaches do 
you, the implementors of BPEL, see as the most likely to demonstrate 
BPEL compliance, portability, and (reasonable) interoperability? Should 
we start trying to spin up an effort (perhaps external) to do this?

    I believe that doing nothing in this domain will do a lot of harm to 
BPEL. Are we at the right stage in BPEL's life to worry about this? Or 
has someone started addressing this issue already?

Cheers,
-Ron