Re: Sample getter/setter take 3

Savas Parastatidis wrote:
> Roberto,
> 
> Thank you for sending this. Please, see a comment bellow.
> 
> [snip]
> 
> 
>>     <ws:operation
>>          name="get_XXX"
>>          style="http://www.w3.org/@@@@/@@/wsdl/style/get-attribute"
>>          pattern="http://www.w3.org/@@@@/@@/wsdl/in-out">
>>       <ws:input name="in" body="wrappers:get_XXX"/>
>>       <ws:output name="out" body="wrappers:get_XXXResponse"/>
>>     </ws:operation>
>>
>>     <ws:operation
>>          name="set_XXX"
>>          style="http://www.w3.org/@@@@/@@/wsdl/style/set-attribute"
>>          pattern="http://www.w3.org/@@@@/@@/wsdl/in-out">
>>       <ws:input name="in" body="wrappers:set_XXX"/>
>>       <ws:output name="out" body="wrappers:set_XXX"/>
>>     </ws:operation>
>>
> 
> What is the difference between the two operations above and the ones
> bellow?
> 
> <ws:operation name="get_XXX"
>               pattern="http://www.w3.org/@@@@/@@/wsdl/in-out">
>    <ws:input name="in" body="wrappers:get_XXX"/>
>    <ws:output name="out" body="wrappers:get_XXXResponse"/>
> </ws:operation>
> 
> <ws:operation name="set_XXX"
>               pattern="http://www.w3.org/@@@@/@@/wsdl/in-out">
>    <ws:input name="in" body="wrappers:set_XXX"/>
>    <ws:output name="out" body="wrappers:set_XXX"/>
> </ws:operation>

On the wire, none: operation styles are transparent to protocol
bindings. On the other hand, language bindings will likely do things
with them. Additionally, other interfaces, e.g. grid or management ones,
may take advantage of these styles to recover the list of attributes
declared by a service and use that information for their own purposes.

> In other words, what would the accompanying language be for the specific
> @style URI? What does such a URI suggest? Is it the case that a
> <blabla>/{get|set}-attribute URI is going to be just a hint to tooling?

We discussed this during the call and Steve Graham is going to send out
a writeup with all the rules. They're going to be similar to the RPC
ones, e.g. if the operation is tagged with the "get-attribute" style
then it MUST satisfy the following conditions: the pattern MUST be
in-out, the input message MUST look like this and the output message
MUST look like that, any number of faults MAY appear, yadda yadda. Plus
some rules to constrain pairs of correlated get/set operations.

> I still don't see the need for this but I guess an optional @style URI,
> with no suggested semantics for a Web Service, is a reasonable
> compromise.

I'm glad to hear that.

Cheers,
Roberto

-- 
Roberto Chinnici
Java Web Services
Sun Microsystems, Inc.
roberto.chinnici@sun.com

Received on Tuesday, 7 October 2003 12:52:15 UTC