Recently, I was attending the OASIS symposium. At the SAP hosted reception of this event, someone was pointing to me that WS-Addressing specs reached recommendation stage in w3c. I paused for a moment and reflected on why I did not share his enthusiasm although I should be happy for getting closure for these specs as a working group member. Unfortunately, a very important part of WS-Addressing family of specifications has not reached that stage yet, and that is the WS-Addressing WSDL binding specification. I am an editor of this specification.

Do you know what I am asked about by our developers most frequently whether it is the Java or ABAP implementation? Well, we have the WS-Addressing and SOAP bindings, but how do we put a markup in WSDL to show that WS-Addressing may/must be engaged? My colleaques whether they are implementing the submission version of WS-Addressing or the newly blessed recommendation have the same problem. What markup to use in WSDL NOW? This is such a practical question. Do you invent your "assertion", use some other vendors proprietary markup or do you use the approach defined in the WSDL binding document, and designate the use of WS-Addressing with the wsaw:UsingAddressing element and the wsaw:Anonymous element?

However, how likely is it that the last approach is not going to change soon? Unfortunately, I could not attend the last face to face meeting, but the debate in the working group appears to be the same.

WSDL Binding document unfortunately targets both WSDL 1.1 and WSDL 2.0. In order to comply with W3C requirements for interoperability, this caused a big headache because the whole specification now lags in order to wait for WSDL 2.0 schedule to go anywhere. Instead of splitting the document into WSDL 1.1 and WSDL 2.0 related bindings where the needs of the community that build upon WSDL 1.1 could be addressed today, this artificial delay was created and the WSDL binding document can not reach the recommendation stage unless the dependent specification, WSDL 2.0 is finalized! As a result, are we in a better place for interop? W3C should have realized this and allowed the specs to be split into two long ago to address the needs of the community.

To confuse the issue, some folks are now questioning whether we need a WS-Policy based markup for WS-Addressing and whether we should "wait" even further. My answer for that is "Come on!". WS-Policy is submitted to the W3C, but using the submitted version as a basis for an assertion could easily be found deficient for a large camp of people who may say that WS-Policy is not a "standard" and hence such an assertion should not be defined, YET. This is again a very misguided argument. The way that the wsaw:UsingAddressing element in the WSDL Binding specification is defined, it is a global element and there is no technical reason why the same element can not be used as a WS-Policy based assertion, today. Therefore, whether you want to use WS-Policy or use the markup for what it is today, it is possible to use the same element in two different contexts. The debate is, thus, just political.

The facility in WSDL Binding document is sufficient enough and should be standardized by the W3C WS-Addressing working group as soon as possible in order to meet the needs of developers who are trying to make service enablement real. I fear that the WSDL binding specification will stay in the CR phase, hence not be a recommendation for a while due to not splitting the specification and thus allowing vendors to do a partial implementation!. This would be very undesirable. WS-Addressing by itself is necessary but not sufficient. Its relationship to WSDL is really needed to be standardized as soon as possible make service enablement of applications real.