Glen,
Thanks for getting the discussion going.
As WSDL is now in CR, and we're trying to get a test suite together,
this is clearly an "as soon as possible" kind of thing, so we're
wondering if people on the Axis2 and Woden teams will have some cycles
in order to get this going in the near term.
Woden has made some pretty significant progress but there are some rather large
work items left including parsing of extension HTTP components, validation of
extension components and MEPs and adding a URI resolution framework. There is
also some work that we will need to do on XmlSchema and may tests we need to
create in order to validate Woden's functionality. I am willing to help,
especially with work that needs to be done to Woden itself. I'm not sure how
much time I'll have in the next couple months to contribute directly to Axis.
At present, Woden does not parse WSDL 1.1
True. We've had some discussion about how to support both WSDL 1.1 and 2.0.
Woden did receive an initial contribution of the WSDL4J source. (AFAIK Axis
using WSDL4J for WSDL 1.1 parsing - correct me if I'm wrong.) The long term
plan may include merging this code with the WSDL 2.0 code and creating one
package that supports WSDL 1.1 and 2.0 but it's too early to say if and when
this will happen. We are of course willing to listen to suggestions and happy
to receive contributions. As it currently stands, I think option 1 makes the
most sense. The WSDL 1.1 - 2.0 conversion utilities do not handle 100% of WSDL
1.1 cases and I don't want to see Axis regress because of an implementation
detail.
I think it's also important to point out that one of Woden's goals is to
answer the W3C WSDL working group's call for implementations. Obviously Woden
can't completely answer the call so we need Axis to be able to achieve this
goal. Also, the plan is for Woden to declare an API. This can't happen until we
receive feedback from clients. Integration with Axis should allow for the type
of feedback we need from an initial adopter.
Getting the dialog going between Axis and Woden is a great first step in
achieving goals for both projects. Thanks for starting the conversation.
Lawrence Mandel
Glen Daniels <[EMAIL PROTECTED] >
03/13/2006 08:32 PM
Please respond to
[email protected]
To [email protected], [email protected]
cc
Subject [axis2] WSDL 2.0 CR work
(Woden folks - please leave the [axis2] prefix on the subject when
replying, as folks use that to filter axis-dev mail into separate folders)
Hi y'all!
I've taken an action from the W3C WSDL working group to begin inquiring
about whether we might be able to get Axis2 integrated with Woden in
time to use that combination as one of the test implementations in order
to prove interoperability of the WSDL 2.0 specs.
As WSDL is now in CR, and we're trying to get a test suite together,
this is clearly an "as soon as possible" kind of thing, so we're
wondering if people on the Axis2 and Woden teams will have some cycles
in order to get this going in the near term.
The test suite we end up with is going to have at least two parts - one
involving parsing WSDL documents, which is pretty much covered by Woden,
and the other involving on-the-wire tests of messages described by WSDL
documents, and that's where Axis2 comes in.
At present, Woden does not parse WSDL 1.1, and WOM does not parse WSDL
2.0... :( Therefore it's not just a matter of unplugging WOM and pluggin
in Woden. So in order to get this working I see several possible
directions we could take:
1) Use Woden to parse WSDL 2.0 documents, and then build a connecting
library which builds WOM objects out of Woden objects. PRO: none of the
rest of the Axis2 code would need to change. CON: the shim layer would
be throwaway code.
2) Add WSDL 1.1 parsing functionality to Woden. PRO: This is where we
really want to be long-term. CON: lots of effort.
3) Switch over to Woden even without WSDL 1.1 parsing functionality, and
use the W3C's WSDL 1.1 -> WSDL 2.0 XSLT in order to do WSDL2Java with
1.1 documents in the meanwhile. PRO: probably the minimum amount of
work. CON: doubtful all WSDL 1.1 documents would work, and there's a
danger we'd fail on some popular services.
Are there others?
We're looking for thoughts, ideas, and ideally people who would be
willing to step up and help with this process in some way. If you reply
please cc both axis-dev and woden-dev.
Thanks,
--Glen
