BPEL, by its nature, requires participation from not only IT, but business analysts. What kind of challenges does that pose?
BPEL is a complex specification. To do it with the right amount of rigor to be a foundation for industry, takes a lot of deep thinking to get into bowels of how it should work with XPath (XML Path Language), WSDL (Web Services Description Language) and XML, for example, and how it should be used by business analysts.
It's exactly true that one of interesting things is scope of participation. Companies and individuals interested in the Web services aspect of BPEL get into the XML and WSDL discussions, and companies predominantly concentrating around business process management and workflow are equally interested from different perspective. A challenge as one of the chairs is to bridge that gap and make sure all of us use the same terminology. BPEL does have some wide applicability. The technical committee is scheduled to meet next week [Tuesday-Thursday]. What's on the agenda?
The agenda is set based on proposals members bring forward and issues they identify. We'll be talking about a few clarifications on how to use XPath, a higher level discussion around whether there should be consideration for sub-processes in BPEL. There will also be a lot of discussion of applications of abstract processes, how they should be used and support that needs to be in the specification.
How is public feedback incorporated into what the TC does?
We don't get a lot of public comment. OASIS is a large organization, and we are one of the larger technical committees. We've got 45 active working members who attend that represent 35 to 40 companies, and that includes all of the big middleware vendors. Many of those companies are implementing BPEL runtime engines [in their products] and many are working at the tooling level and how to provide graphical interfaces that will support BPEL. On the other hand, a number of individuals are in the business analyst role. Can you talk about BPEL's importance among the vast array of Web services specifications already approved or under construction?
Speaking as an IBM representative for a second, the origins of BPEL come from work done at IBM and Microsoft on WSSL (Web Services Security Language) and XLANG (the XML business process language used in Microsoft's BizTalk Server). With the marriage of those two technologies, BPEL has taken a giant leap forward.
It provided BPEL with the critical capabilities that allow Web services applications to go up a notch in terms of their business value. Basic Web services are a great concept, but a stateless one-on-one interaction. With BPEL, you take activities as part of a business process, express them as a Web service, then orchestrate Web services so you have control over the full process. It's compelling from a business process point of view and from a process management point of view. What is driving all the interest in BPEL? Is it IT's growing interest in service orientation?
Sure, that's part of it. The timing of BPEL is well suited. It's gotten to the point where the foundation of Web services is understood and adopted. Companies have gotten through the investigative and proof-of-concept phases and are thinking about integrating the technology into real work and how to use Web services.
If you take the classic example of booking a business trip online, you're communicating from a client company through a travel agent to an airline, hotel or car rental agency. You need to be able to orchestrate those steps and determine relationships between them.
From a standardization perspective, the beauty of BPEL, once you work through the details, is that it has the potential to offer a degree of interoperability at the process level, describe process and use it in different runtime infrastructures. It opens the possibility for standardizing processes as a deliverable expressed in BPEL. As the TC deliberates, any surprises or sticking points you can share?
What we're working on with BPEL is the core foundation for processes and process management, and taking into consideration any ideas for extensions [to the specification], most of which we are deferring. We are deferring because there is a great deal of work to be done to get standardized. Any interesting ideas, we are logging for future consideration.
The sticking points revolve mainly around how it has taken members some time to fully appreciate the aspects of the spec and the relationships between the components of BPEL.
FEEDBACK: Where do you rank BPEL among the Web services standards?
Send your feedback to the SearchWebServices.com news team.
The technical committee is chartered to complete three deliverables: specify the core elements and functionality of BPEL; an extension specification for business protocol description; and an extension specification for executable process description. Can you tell me which, if any, have been completed and provide a progress report on those that remaining?
The original charter was written with the input document in mind. It has three sections and these were referenced as three deliverables but there hasn't been a plan to actually split them into different documents. We've been working on all three sections in parallel and the final specification will include all the content. There's been a suggestion to rework the spec for readability which might result in a different structure where there wouldn't even be three parts. That will come up for consideration during the final editing once we've dealt with the technical issues and clarifications.