RE: [getf] Proposal for Web-friendly representation of RPC's in SOAP

>> If it is a representation, then any transform is allowable I would 
imagine.

That's what I intended when I proposed that the outer element be viewed as 
a "constructor".  It is not a call to an arbitrary method:  it defines the 
way (I.e. it's QName is associated with a specification) that the 
arguments can be used to establish the state of the resource to which the 
PUT is directed.  Indeed, this is intentionally very similar to a non-RPC 
document-oriented PUT, which I would also support.  The only significance 
of it being conformant to RPC is that bindings to certain programming 
language structures will be relatively easy, if that's how you choose to 
structure your implementation at the endpoint. 

Concrete example:

Consider two different PUT's to the same resource (same URI) which stores 
a string of English language text.  For whatever reasons, this resource 
exposes two constructors:  setFromEnglish and setFromFrench.  The 
specification for the latter says:  "an automatic machine translator will 
translate the supplied string from French to English and use the result to 
set the state of the resource.  So, the following produce the same value 
for the resource:

PUT:
<soap:Envelope>
        <soap:body>
                <createFromEnglish>
                        Hello
                </createFromEnglish>
        </soap:body>
</soap:Envelope>

PUT:
<soap:Envelope>
        <soap:body>
                <createFromFrench>
                        Bonjour
                </createFromFrench>
        </soap:body>
</soap:Envelope>

Each takes a representation that completely establishes the state of the 
resouce, but using different conventions. Why are these not perfectly 
appropriate uses of PUT?  That was my intention.

------------------------------------------------------------------
Noah Mendelsohn                              Voice: 1-617-693-4036
IBM Corporation                                Fax: 1-617-693-8676
One Rogers Street
Cambridge, MA 02142
------------------------------------------------------------------







Mike Dierken <mike@dataconcert.com>
Sent by: xml-dist-app-request@w3.org
05/30/2002 09:34 PM

 
        To:     "'Mark Baker'" <distobj@acm.org>, xml-dist-app@w3.org
        cc:     (bcc: Noah Mendelsohn/Cambridge/IBM)
        Subject:        RE: [getf]  Proposal for Web-friendly representation of RPC's in SOAP





> -----Original Message-----
> From: Mark Baker [mailto:distobj@acm.org] 


> 
> PUT's definition doesn't allow this wiggle room; what's in 
> the body is the desired state of the resource identified by 
> the Request-URI. Period.
Is the entity body merely a representation of the desired state of the
resource? Or a 'canonical' form?
If it is a representation, then any transform is allowable I would 
imagine.

Received on Friday, 31 May 2002 00:21:03 UTC