RE: Revised proposed text for intermediary

>
>I don't think that the definition says that the SOAP intermediary has
>to be explicitely targeted, only that they can be.
>
>> If we are not relaxing the definition, then I would s/may/must/
>
>I used "may process" thinking about the case of an active intermediary
>which would end up inspecting the message and making the decision of
>not doing anything to it and forwarding it as received. However,
>"processes" works as well in the end: after all, parsing and making
>the decision of not doing anything is processing.

According to the section 2 of the 1.2 spec, "A SOAP node receiving a
SOAP
message MUST (their emphasis) perform processing according to the SOAP 
processing rules". The processing rules then takes over and forces the
SOAP 
node to process in the role which that node operates. 

So I am not certain if 'inspecting the message and making the decision
of not 
doing anything' is an option. It seems that the decision to do or not do
something 
is completely deterministic and is based on the fact that the SOAP
message arrives 
at the SOAP (intermediary) processor and whether any headers contain the
role attribute 
with value 'next'.


>
>Does the revised text ([2] with the modification [3]) address your
>concerns?

The section about active intermediaries in the 1.2 spec may answer my
confusion 
above and be the formalism which an 'inspection' only of a SOAP message
can occur.
Is that your interpretation? If this is the case - then a SOAP Active
Intermediary
can have no headers targeted at it, and can insert, remove, or edit
headers at will. 


>
>Regards,
>
>Hugo
>
>  2. http://lists.w3.org/Archives/Public/www-ws-arch/2003Jul/0107.html
>  3. http://lists.w3.org/Archives/Public/www-ws-arch/2003Jul/0110.html
>-- 
>Hugo Haas - W3C
>mailto:hugo@w3.org - http://www.w3.org/People/Hugo/
>

Received on Thursday, 17 July 2003 11:41:41 UTC