Additional Blogs by SAP
cancel
Showing results for 
Search instead for 
Did you mean: 
kevinliu
Product and Topic Expert
Product and Topic Expert
0 Kudos

After more than two years diligent work, the W3C WSD working group has finally put together a last call draft for WSDL2.0. You may have noticed the W3C announcement last week (see [6]) while I was on vacation, but I would like to bring this Coverpage news to your attention again with a few key points:

First of all, WSDL2.0 is NOT just a rubber stamp for WSDL1.1. In addition to clarification and corrections to the WSDL1.1 spec, WSDL2.0 has incorporated quite a few re-designs of key features. For example,

- wsdl11:message construct is removed, messages are directly defined using XML schema or other type languages.
- RPC style is demoted as a "feature".
- Added support for interface inheritance/aggregation.
- Added extensible message exchange pattern support.
- All wsdl constructs have been made extensible via both elements and attributes.
- portType is renamed to interface, port is renamed to endpoint...

The redesign will hopefully make your future work with WSDL easier, but it does have some side effect - wsdl2.0 is not really backward compatible with 1.1 (just like SOAP 1.2 is not backward compatible with 1.1), so you need to analyze your existing wsdls and consider how to migrate to the new version.

Secondly, while you review the draft, it's worth noting that the last call draft still contains a few controversial proposals.

-- Inclusion of features and properties is objected to by IBM, Microsoft, and SAP(see [1]). SAP believes Web services feature and property support is very important, but WSDL is the wrong place to address it. It deserves its own specification to provide a comprehensive solution. In fact, W3C is organizing a workshop on WS-policy and related proposals. As a result of that workshop, most likely a W3C working group will be set up to address feature and property support.

The functionality provided in WSDL2.0 overlaps with WS-Policy and unnecessarily complicates the WSDL language. In addition, the lack of a composition framework really limits the usefulness of WSDL F&P (see [2] for the objection from Oracle, SUN and a few others about "compositor"). Given the amount of work involved to develop a useful composition framework, we don't believe WSDL is the right place for such a complete solution.

If you know of any customers who agree with our objection, please encourage them to express their concerns to w3c. (via the WSD mailing list, see [4])

-- Microsoft, IBM, and TIBCO have also filed an objection to requiring unique GEDs or required features to distinguish operations (see [3])

All the drafts are downloadable from W3C (see [5]). Please note that it's a "last call" for comments, so get serious if you care about WSDL and submit your comments to me or to W3C directly via http://www.w3.org/2002/ws/desc/4/lc-issues/ before 4 October 2004...

[1] http://lists.w3.org/Archives/Public/www-ws-desc/2004Jul/0395.html

[2] http://lists.w3.org/Archives/Public/www-ws-desc/2004Jul/0371.html

[3] http://lists.w3.org/Archives/Public/www-ws-desc/2004Jul/0376.html

[4] http://www.w3.org/2002/ws/desc/4/lc-issues/

[5] http://www.w3.org/2002/ws/desc/

[6] http://www.w3.org/News/2004