W3C home > Mailing lists > Public > www-ws-desc@w3.org > June 2005

RE: Additional HTTP and safety primer example and text

From: Liu, Kevin <kevin.liu@sap.com>
Date: Fri, 17 Jun 2005 20:16:31 +0200
Message-ID: <3470F33FF8ED12498D07F3A9651AA18E207E30@uspale20.pal.sap.corp>
To: <paul.downey@bt.com>, <dorchard@bea.com>, <www-ws-desc@w3.org>

+ 1. 

The related part 2 section [1] is consistent with Paul's amendament as
saying:

<quote>
The actual value of the whttp:method attribute information item, if
present; otherwise, the actual value of the whttp:methodDefault
attribute information item, as defined in 6.5 Specifying the Default
HTTP Method; otherwise, if a {safety} property as defined in 3.1
Operation safety is present on the bound Interface Operation component
and has a value of "true", the value "GET"; otherwise, it is an error. 
</quote>  

Also note as the TAG finding in [2] pointed out there are situations
that GET is not a natural choice even if the operation is marked safe.
Two common examples quoted are when sensitive data is passed as
parameters nd when the parameter size is just too large to be encoded in
a URL. 

[1]
http://dev.w3.org/cvsweb/~checkout~/2002/ws/desc/wsdl20/wsdl20-adjuncts.
html#tab_HTTP_Operation_Mapping

[2] http://www.w3.org/2001/tag/doc/whenToUseGet.html

Dave, if you have no objection, I will add the primer text in the spirit
of "defaulting". 

Best Regards,
Kevin
  

> -----Original Message-----
> From: paul.downey@bt.com [mailto:paul.downey@bt.com] 
> Sent: Friday, Jun 17, 2005 2:36 AM
> To: Liu, Kevin; Liu, Kevin; dorchard@bea.com; www-ws-desc@w3.org
> Subject: RE: Additional HTTP and safety primer example and text
> 
> my understanding is safety can be used to provide a default verb which
> may be overridden by the HTTP binding. i.e:
>  
> """
> 2) The HTTP binding would take a dependency upon the safe 
> attribute, and
> when the safe property is set to true, the HTTP method would 
> default to
> GET, though this can be overridden when other considerations 
> (e.g. data
> not easily serialized in a URI) apply.  This satisfied those who felt
> safety was important enough to not only remain in the family of
> Recommendations, but to have a real effect on the HTTP binding.
> """
>  
> http://lists.w3.org/Archives/Public/www-tag/2005Jun/0009.html
>  
> I suggest changing Dave's wording of 'will set the method' to 
> 'will by default
> set the method' or some such
>  
>  
> 
> 	-----Original Message----- 
> 	From: www-ws-desc-request@w3.org on behalf of Liu, Kevin 
> 	Sent: Thu 6/16/2005 11:49 PM 
> 	To: Liu, Kevin; David Orchard; WS-Description WG 
> 	Cc: 
> 	Subject: RE: Additional HTTP and safety primer example and text
> 	
> 	
> 	David,
> 	 
> 	I have changed the primer to reflect the changes we 
> agreed upon below, except one item,
> 	 
> 
> 	"Addition to 6.7
> 
> 	 
> 
> 	wsdlx:safe="true"
> 
> 	on the interface operation element.  The HTTP Binding 
> will set the method to GET if wsdlx:safe="true"
> 	"
> 
> 	 
> 
> 	when tried to make this change, I realized that it 
> contradicts with the status quo of section 6.7 which says: 
> 
> 	 
> 
> 	"Although the wsdlx:safe attribute of an interface 
> operation indicates that the abstract operation is safe, it 
> does not automatically cause GET to be used at the HTTP level 
> when the binding is specified. The choice of GET or POST is 
> determined at the binding level: "
> 
> 	 
> 
> 	 
> 
> 	As I stated in my initial response, such change 
> involves semantic changes to wsdlx:safe. It needs the 
> approval of the group, and should be reflected in the core 
> specs first before included in the primer.
> 
> 	 
> 
> 	Best Regards,
> 	Kevin
> 	  
> 
> 	 
> 	
> 	
> 
>   _____  
> 
> 		From: www-ws-desc-request@w3.org 
> [mailto:www-ws-desc-request@w3.org] On Behalf Of Liu, Kevin
> 		Sent: Thursday, Jun 16, 2005 12:13 PM
> 		To: David Orchard; WS-Description WG
> 		Subject: RE: Additional HTTP and safety primer 
> example and text
> 		
> 		
> 		David, 
> 		 
> 		Thanks a lot to thinking this through.  
> Actually at this moment, I am just looking into how to 
> reflect the changes of safety in the primer, your proposal 
> saves me a lot of time:))
> 		 
> 		I will incorporate pretty much all your 
> propsoals except a few point that need clarification. See my 
> comments below.
> 
> 		Best Regards,
> 		Kevin
> 		  
> 
> 		 
> 		
> 		
> 
>   _____  
> 
> 			From: www-ws-desc-request@w3.org 
> [mailto:www-ws-desc-request@w3.org] On Behalf Of David Orchard
> 			Sent: Wednesday, Jun 15, 2005 9:59 PM
> 			To: WS-Description WG
> 			Subject: Additional HTTP and safety 
> primer example and text
> 			
> 			
> 
> 			Section 2.4 and 5.1 example should 
> change safe to wsdlx:safe
> 			[Kevin]  will do. we also need to 
> change example 2.1
> 
> 			 
> 
> 			I suggest 5.4.1 needs some text to 
> describe how wsdlx:safe can be used.  
> 			[Kevin] yes. 
> 
> 			 
> 
> 			there are already 3 paragraphs in 5.4.1 
> explaining safety attribute. I am thinking the only change 
> needed is to reflect the fact that safe is not an optional 
> attributes of interface/operation any more, but a global 
> attribute that can be used with interface/peration. what do you think?
> 
> 			 
> 
> 			 My suggestion: The wsdlx:safe 
> attribute may be used in bindings.  The HTTP binding uses a 
> "true" setting of wsdlx:safe to indicate that HTTP GET is the 
> operation, simplifying the HTTP Binding.
> 			[Kevin] is this adopted semantic for 
> the wsdlx:safe attribute? If so, I have no problem adding it 
> to the primer. But I don't want to introduce new semantic via 
> the primer.
> 
> 			 
> 
> 			Example 6.2 needs 
> type="http://www.w3.org/2005/05/wsdl/http" in the binding.
> 			[Kevin] yes, will add.  
> 
> 			 
> 
> 			Addition to 6.7
> 
> 			 
> 
> 			wsdlx:safe="true"
> 
> 			on the interface operation element.  
> The HTTP Binding will set the method to GET if wsdlx:safe="true"
> 			[Kevin] yes, will do 
> 
> 			 
> 
> 			 
> 
> 			I suggest a new section 6.8
> 
> 			 
> 
> 			6.8 Safety and the HTTP Binding
> 
> 			In the GreatH interface definition 
> shown in example 2.4, the wsdlx:safe attribute = "true".  The 
> HTTP binding will use this value to set the http method to 
> GET, meaning that methodDefault on binding or method on 
> binding operation do not need to be set for HTTP GET.  The 
> HTTP Binding can be simplified to:
> 
> 			 
> 
> 			<binding name="reservationHTTPBinding"
> 
> 			      interface="tns:reservationInterface"
> 
> 			      
type="http://www.w3.org/2005/05/wsdl/http" >
> 
> 			 
> 
> 			    <operation ref="tns:opCheckAvailability"
> 
> 			        whttp:location="{checkInDate}"/>
> 
> 			  </binding>
> 
> 			 
> 
> 			The binding operation has used the 
> wsdlx:safe attribute to generate the HTTP GET operation.
> 			[Kevin]  if this is adopted semantic 
> for the wsdlx:safe attribute, i can certainly add this, but 
> would probably add as a subsection of 6.7
> 
> 			 
> 
> 
Received on Friday, 17 June 2005 18:16:56 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:58:36 GMT