RE: Version attribute for WSDL

A big plus 1 to the statements about contracts versus implementation and
substitutability.  The implementation and the contract are different.  There
may be many interfaces to the same code (forwards/backwards compatibility)
and there may be 1 interface to many different sets of code.

However, it's more than just syntactic evolution.  the utility of version +
ns is that the author of the WSDL file can make an assertion about
versioning and compatibility.  The same syntactic change (say adding an
optional element) may to one WSDL author be an incompatible change - the
receivers are required to know about the element - whereas another WSDL may
say it is a compatible change - that receivers are not required to know
about the element.

And further, there may be semantic changes though the syntax remains the
same.  If I change my purchase order software in some incompatible way, such
as billing on ship to billing after 30 days, the author may decide that that
is an incompatible semantic change, and old clients cannot continue to use
the old purchase order interface.

Compatibility is really in the eye of the service provider.  The combination
of namespace name + version (structure/integer/uri/whatever) gives them full
fidelity needed for them to make assertions about what senders and receivers
are required to know and use.

We already have the basic with the use of the target namespace and wsdl
component identifiers, that is authors can decide to change the component
identifiers in a way they choose.  They can keep the same names or they can
change them when versioning.

Cheers,
Dave

> -----Original Message-----
> From: www-ws-desc-request@w3.org [mailto:www-ws-desc-request@w3.org]On
> Behalf Of Jeff Mischkinsky
> Sent: Friday, February 13, 2004 2:31 PM
> To: Tom Jordahl; 'www-ws-desc@w3.org'
> Subject: RE: Version attribute for WSDL
>
>
>
> I like the way the discussion is going (including the
> following emails with
> suggestions how to handle revisions using uri's, etc.) Let's keep it
> simple. Just covering the simple obvious cases will be very helpful.
>
> The one thing I think we need to be very clear on is what we mean by
> phrases like "continues to work with" etc.
>
> It is my belief that we are not making statements about the
> semantics that
> are implied by describing two services as being related by
> the "version of"
> relationship. Rather we are allowing the designers to make claims and
> provide assurances about the syntactic evolution of the interface
> types/signatures, i.e. the infrastructure isn't going to fall
> over when an
> old client using the old signature talks to a newer
> (compatible) version.
> Essentially we are defining the substitutability rules.
>
> Obviously what the implementation code that lies behind the
> web service
> facade actually does is not defined in WSDL. In fact, when
> you get right
> down to it, it is possible that 2 successive uses of the
> "same" interface
> will use 2 different implementations if the implementations
> are swapped
> without changing the signatures. We even call those the same
> version, even
> though what happens may be different.
>
> jeff
>
> At 07:05 AM 2/13/2004, Tom Jordahl wrote:
>
>
> >I guess I understand the desire to have "real" versioning
> support in WSDL
> >2.0.  I do too. But my proposal came out of the F2F, where
> we had a long,
> >and I believe fruitless, discussion about all of this.
> >
> >I do not believe we can have a section in our specification
> about versioning
> >and say "if you want versioning, change the namespace".  With a small
> >addition to the syntax, we can give users some help in
> managing change in
> >their web services.
> >
> >I am willing to apply semantics to the version attribute if
> this group
> >thinks that they can move forward in a productive way.  How
> about these
> >changes as a straw man for discussion:
> >
> >  - The version attribute is part of the infoset (a.k.a. the
> component model)
> >
> >  - The version attribute has type xsd:positiveInteger
> >
> >  - The version attribute has a default value of 1.
> >
> >  - The version attribute indicates a "minor" revision of
> the definition or
> >    interface. Specifically, a "minor" revision indicates
> that a client using
> >    a WSDL with a version attribute less-than the current
> value is expected
> >    to continue to function.
> >
> >  - When an interface extends another interface, the version
> attribute of the
> >    interface is NOT inherited - it must be explicitly set
> on the interface,
> >    and if is not, the interface has the default version
> attribute (1).
> >
> >Example 1: Version 1 of my interface has two operations. I
> release a new
> >WSDL that adds a third operation, and change the version
> attribute to 2.
> >Clients who are using the previous version of the WSDL
> continue to function.
> >
> >Example 2: My WSDL has a purchase order type defined and a
> target namespace
> >of "http://example.org/myservice".  I change my purchase
> order to include
> >several new elements and rename some of the old ones.  Since
> this change
> >will break compatibility, I change the target namespace to
> >"http://example.org/myservice/v2".  My service can now
> easily tell the
> >difference between clients that are using the original WSDL
> and the new one.
> >
> >
> >--
> >Tom Jordahl
> >Macromedia Server Development
> >
> >-----Original Message-----
> >From: Jeff Mischkinsky [mailto:jeff.mischkinsky@oracle.com]
> >Sent: Friday, February 13, 2004 2:56 AM
> >To: Tom Jordahl; 'www-ws-desc@w3.org'
> >Subject: RE: Version attribute for WSDL
> >
> >At 12:41 PM 2/12/2004, Tom Jordahl wrote:
> >
> >
> > >David,
> > >
> > >We wouldn't say anything like this about the version attribute.
> > >      "..it has no semantics.."
> > >
> > >So David can tell his WSDL consumers that he uses this attribute to
> >indicate
> > >compatible versions of the same WSDL file.  And I can tell
> my users that
> > >version 1 does not equal version 2.  But as WSDL spec
> authors we don't have
> > >to take a stand on how this is done.
> > >
> > >Isn't that nice?  We don't have to fight about what it means.
> >
> >and the point of "standardizing" this would be?
> >(in this case I'm using the word standardize in its loosest most
> >meaningless sense :-)
> >
> >jeff
> >
> >
> > >--
> > >Tom Jordahl
> > >Macromedia Server Development
> > >
> > >-----Original Message-----
> > >From: David Orchard [mailto:dorchard@bea.com]
> > >Sent: Thursday, February 12, 2004 2:32 PM
> > >To: paul.downey@bt.com; vbp@hp.com; tomj@macromedia.com;
> www-ws-desc@w3.org
> > >Subject: RE: Version attribute for WSDL
> > >
> > >I'm interested in the version attribute for identifying
> versions within
> > >"compatible" definitions.  I would like to have our spec
> say explicitly
> > >that.  I am strongly strongly opposed to using a version
> attribute for
> > >identifying different incompatible versions.  That's what
> namespaces and
> > >URIs are for.
> > >
> > >Some off-the-cuff suggestions for the wording:
> > >
> > >"The version attribute identifies a particular version of
> the definitions,
> > >that is compatible with all other versions with the same
> targetnamespace.
> > >It SHOULD not be used to identify incompatible definition
> versions."
> > >
> > >Cheers,
> > >Dave
> > >
> > > > -----Original Message-----
> > > > From: www-ws-desc-request@w3.org
> [mailto:www-ws-desc-request@w3.org]On
> > > > Behalf Of paul.downey@bt.com
> > > > Sent: Thursday, February 12, 2004 9:02 AM
> > > > To: vbp@hp.com; tomj@macromedia.com; www-ws-desc@w3.org
> > > > Subject: RE: Version attribute for WSDL
> > > >
> > > >
> > > >
> > > > I believe the version value is useful information for when the
> > > > interface has been compatibly changed within the same namespace.
> > > >
> > > > +1 Tom's proposal, i can't see any harm and it could be useful
> > > > as a building block for a mechanism for relating an interface
> > > > version to other versions, akin to the 'previous', 'this' and
> > > > 'latest' version URLs on W3C publications.
> > > >
> > > > Paul
> > > >
> > > >
> > > > -----Original Message-----
> > > > From: www-ws-desc-request@w3.org
> [mailto:www-ws-desc-request@w3.org]On
> > > > Behalf Of Vambenepe, William N
> > > > Sent: 12 February 2004 16:53
> > > > To: Tom Jordahl; www-ws-desc@w3.org
> > > > Subject: RE: Version attribute for WSDL
> > > >
> > > >
> > > >
> > > >
> > > > Thanks Tom for the proposal. I could live with this attribute on
> > > > <definitions> but I really don't like it on <interface>. As Glen
> > > > eloquently explained at the F2F, a different interface
> should use a
> > > > different QName. What does it mean for a binding to reference an
> > > > interface if there are dozens of "versions" of this
> interface. Can I
> > > > have a binding for only a certain version of an
> interface? I know we
> > > > don't have to answer this since we "define no semantic" but
> > > > that doesn't
> > > > make the problem go away.
> > > >
> > > > William
> > > >
> > > >
> > > > > -----Original Message-----
> > > > > From: www-ws-desc-request@w3.org
> > > > > [mailto:www-ws-desc-request@w3.org] On Behalf Of Tom Jordahl
> > > > > Sent: Thursday, February 12, 2004 6:13 AM
> > > > > To: 'www-ws-desc@w3.org'
> > > > > Subject: Version attribute for WSDL
> > > > >
> > > > >
> > > > >
> > > > >
> > > > > In fulfillment of my action item received at the January F2F,
> > > > > here is a
> > > > > proposal to add a version attribute to WSDL to aid in the
> > > > > versioning of WSDL
> > > > > documents and interfaces.
> > > > >
> > > > > I propose that an attribute with the name "version"
> be added to the
> > > > > <definitions> element of WSDL.  This attribute is for user
> > > > > convenience, and
> > > > > the specification would define no semantics for it,
> > > > > specifically the value
> > > > > of this attribute would NOT be included in the infoset.
> > > > > However, it is
> > > > > expected that WSDL authors and consumers can use this
> > > > attribute, when
> > > > > present, to differentiate between different revisions of a
> > > > > WSDL document.
> > > > >
> > > > > Example:
> > > > >
> > > > > <definitions version="1" targetNamespace=http://sample.org/>
> > > > > ...
> > > > > </definitions>
> > > > >
> > > > > This proposal is modeled after the version attribute of XML
> > > > > Schema, see
> > > > > section 3.15.2 in Part 1 of the XML Schema specification:
> > > > >   http://www.w3.org/TR/xmlschema-1/#Schemas
> > > > >
> > > > > In our specification, section 2.1.2 would be updated to
> > > > > include the new
> > > > > attribute:
> > > > >
> > > > > 2.1.2 XML Representation of Definitions Component
> > > > >
> > > > > <definitions
> > > > >       targetNamespace="xs:anyURI"
> > > > >       version = "xs:token"? >
> > > > >   <documentation />?
> > > > >   [ <import /> | <include /> ]*
> > > > >   <types />?
> > > > >   [ <interface /> | <binding /> | <service /> ]*
> > > > > </definitions>
> > > > >
> > > > >
> > > > > Additionally, I propose that a similar version attribute be
> > > > > added to the
> > > > > <interface> element of WSDL. This attribute would mirror the
> > > > > definitions
> > > > > attribute.  Again, this would be for user convenience, and
> > > > > the specification
> > > > > would define no semantics for it, specifically the value of
> > > > > this attribute
> > > > > would NOT be included in the infoset.  WSDL authors and
> > > > > consumers could use
> > > > > this attribute, when present, to differentiate between
> > > > > different revisions
> > > > > of an interface.  In particular, this would enable a
> consumer of the
> > > > > document to know explicitly when an interface they are using
> > > > > has changed.
> > > > >
> > > > > Example:
> > > > > <definitions>
> > > > >   <interface name="myInterface" version="alpha17">
> > > > >     ...
> > > > >   </interface>
> > > > > </definitions>
> > > > >
> > > > >
> > > > > 2.2.2 XML Representation of Interface Component
> > > > > <definitions>
> > > > >   <interface
> > > > >         name="xs:NCName"
> > > > >         extends="list of xs:QName"?
> > > > >         styleDefault="xs:anyURI"?
> > > > >         version = "xs:token"? >
> > > > >     <documentation />?
> > > > >     [ <operation /> | <feature /> | <property /> ]*
> > > > >   </interface>
> > > > > </definitions>
> > > > >
> > > > >
> > > > > --
> > > > > Tom Jordahl
> > > > > Macromedia Server Development
> > > > >
> > > > >
> > > >
> > > >
> >
> >Jeff Mischkinsky                      jeff.mischkinsky@oracle.com
> >Consulting Member Technical Staff     +1(650)506-1975
> >Director, Web Services Standards      500 Oracle Parkway M/S 4OP9
> >Oracle Corporation                    Redwood Shores, CA 94065
>
> Jeff Mischkinsky                      jeff.mischkinsky@oracle.com
> Consulting Member Technical Staff     +1(650)506-1975
> Director, Web Services Standards      500 Oracle Parkway M/S 4OP9
> Oracle Corporation                    Redwood Shores, CA 94065
>
>

Received on Friday, 13 February 2004 18:01:38 UTC