See also: IRC log
Present:
Steve Graham GGF
William Vambenepe Hewlett-Packard
Umit Yalcinalp Oracle
David Snelling GGF
Regrets:
Scribe: William
Minutes
of 07/14 telecon approved.
Action item
from Philippe regarding IP statement from GGF still pending.
David Snelling has made an IP statement regarding
attributes from Fujitsu.
<WilliamV>
overview of steve's postings over the we
... defines 4 contexts
... further alignment of the proposal
... clarification of details
... mapping to requirements
... mapping to OGSI needs
... those contexts are refered to as a, b, c and d
... participants: david snellig,
steve, umit, william
... minutes from previous meeting approved
... steve, umit: there
needs to be an operation to access attributes
... william: the message needs to be well-defined but
this doesn't necessarily mean that we need a WSDL operation in the WSDL of all
services that have attributes
... steve: if an operation needs to be defined, where
is this presented? implicit? explicit?
... umit: different opinions
in oracle
... william: similarity between this situation and
the portType/binding situation in WSDL 1.1
... William: <description of proposal>
... david: this is intresting,
it couldn't be done using a WSDL itnerface with
option (a) because you would need to know the name of attributes when defining
the access emchanism
<sgg>
<interface name="someIntf">
... <attribute name="a1" type="xsd:string">
<WilliamV>
umit, william:
you need another abstraction layer. for operations:
operation->binding, for attributes: attribute->accessMethod->binding4accessMethod
<sgg>
... <operation name="getA1" ... />
... </interface>
... <binding type="someIntfBinding">
... <attributeBinding>
... <get = "getA1"/>
... </attributeBinding>
... </binding>
... ??
<WilliamV>
umit: actual WSDL doesn't have to be the same as the
client WSDL (client WSDL could have some added "implicit" operations)
... umit: from client perspective, implicit
operations look like they exist
... steve: 3 approaches to attribute access:
classical getXX/setXX, weak-typed get(name),
set(name, value), and finally amore sophisticated access mechanism
<sgg>
<interface name="sameIntf">
... <attribute name="a1" type="xsd:string">
... </interface>
... <binding name="b1" type="sameIntf">
... <attributeBinding style="http://www.w3.org/wsdl/attributeAccess/optionA"
/>
... </binding>
<WilliamV>
william, umit:
does attributeBinding belong in the binding element
or outside of it?
... steve: this is a good compromise between implicit
operations and explicit portTypes
... discussion now on which access methods to provide in wSDL
1.2
... steve had 2 questions about inheritence
... plan for F2F: umit, steve
(and maybe david and william)
meet wendesday morning at 9
... umit, steve, william meet wednesday evening
... to prepare the F2F presentation