Re: [ws-rx] My vote on the current RX specs ballot

[Paul Fremantle <paul@wso2.com>]

Anish

I personally agree with your reading that this is a product decision. 
Apache products aim to conform to specs closely, but also allow 
interoperable usage in other situations that may go beyond those specs.

A recent example is that we have added support for the RM1.0 replay 
model to interoperate with .NET, even though this requires implementing 
a number of algorithms that go beyond the documented RM 1.0 spec.

However, in this case before us it is my belief that it is the TC's 
intent that WSRMP may be used with either WSP1.2 or WSP1.5CR. Therefore 
I am not averse to us trying to express that in the spec. If that is not 
the TC's intent then I think it should be stated.

Paul

Anish Karmarkar wrote:
> Paul Fremantle wrote:
>> Marc
>>
>> I certainly was of the opinion on the call that this wasn't the TC's 
>> intention, and nor was so literal a reading of the spec.
>>
>> My understanding is that the WSRMP usage of WSP is confined to the 
>> subset of features that is the same from 1.2 to 1.5 CR.
> 
> Does that mean I can't use ignorable (or some other 1.5-specific 
> feature), which is specific to 1.5?
> 
> I agree that WSRMP spec does not use any feature that is specific to 
> 1.5, but WSRMP assertions do not exist in a vacuum. They are used in a 
> particular framework.
> 
>> Certainly I didn't expect that Apache would confine itself to using 
>> WSRMP1.1 with only WSP1.5.
>>
> 
> But isn't that a product decision rather than a spec decision? For 
> example, Apache SOAP supported not just XML Schema REC NS but REC + two 
> additional intermediate (CR and PR, I think) NSs.
> 
> -Anish
> -- 
> 
>> Paul
>>
>> Marc Goodner wrote:
>>> I filed this as a comment with my ballot. Based on some private 
>>> messages I have received I think it is worth posting this to the 
>>> list. It seems that my reading of the change to the RM Policy spec 
>>> may be more literal than the TC’s intent. I’m posting this here with 
>>> the intent of seeing if people disagree or not with my reasoning with 
>>> respect to whether or not the RM Policy assertion can be used with 
>>> the Policy 1.2 namespace even though our reference is now solely to 
>>> Policy 1.5. I am especially interested in how any interpretations 
>>> contrary to my own are supported by what is written in the 
>>> specification.
>>>
>>>  
>>>
>>> Regards,
>>>
>>> Marc g
>>>
>>>  
>>>
>>> The late change from referencing WS-Policy 1.2 to WS-Policy 1.5 in RM 
>>> Policy required that I abstain from the current ballot for the RX TC 
>>> specs.
>>>
>>>  
>>>
>>> This change was not technically necessary as the RM Policy assertion 
>>> does not use any features of WS-Policy 1.5 alone and the assertion is 
>>> backwards compatible with WS-Policy 1.2. The change does however mean 
>>> that Microsoft cannot use the revised RM Policy assertion and be 
>>> compliant to this specification until we also support the final 
>>> version of WS-Policy 1.5.
>>>
>>>  
>>>
>>> I believe that the delay caused by waiting for WS-Policy 1.5 to be 
>>> completed is going to hurt the market adoption of WS-RM and WS-RM 
>>> Policy.
>>>
>>>  
>>>
>>> It is regretful that the TC made this change at the last minute 
>>> thinking of the references alone rather than considering the change's 
>>> impact on market adoption of the specifications.
>>>
>>>  
>>>
>>
> 

-- 
Paul Fremantle
VP/Technology and Partnerships, WSO2
OASIS WS-RX TC Co-chair

http://bloglines.com/blog/paulfremantle
paul@wso2.com
(646) 290 8050

"Oxygenating the Web Service Platform", www.wso2.com