RE: [ws-rx] Concrete proposal for PR021

[Christopher B Ferris <chrisfer@us.ibm.com>]

Gil,

I section 3.8

change:

"The RM Destination
SHOULD process any AckRequested header
blocks that may be 
included in any
message it receives."

to:

"The RM Destination
SHOULD process AckRequested header
blocks that are included in 
messages it receives."

The change above
removes the 'may' which even though uncapitalized, still conveys
the RFC2119 semantic,
which I don't believe is apropriate in this context. 

Similarly, the same change applies to
section 3.9

change:

The RM Source SHOULD
process any SequenceAcknowledgement
header blocks that
may be included in any message it receives.

to:

The RM Source SHOULD
process SequenceAcknowledgement
header blocks that
are included in messages it receives.

Cheers,

Christopher Ferris
STSM, Software Group Standards Strategy
email: chrisfer@us.ibm.com
blog: http://www.ibm.com/developerworks/blogs/page/chrisferris
phone: +1 508 377 9295

"Gilbert Pilz" <gpilz@bea.com> wrote
on 01/06/2007 07:18:27 PM:

> Fair enough, though I have to quibble a bit with your wording (do
> 'end'points send messages?). Attached is a variant that includes a
> revised version(s) or your wording in section 3.2 and changes the
> MUST in 3.8 and 3.9 to a SHOULD.
>  
> - gp
>  
> 
> From: Doug Davis [mailto:dug@us.ibm.com] 
> Sent: Saturday, January 06, 2007 7:05 AM
> To: ws-rx@lists.oasis-open.org
> Subject: Re: [ws-rx] Concrete proposal for PR021

> 
> Sections 3.8 and 3.9: Additional statements of the pattern "The
RM 
> [Source | Destination] MUST detect and process any [*] header blocks
> that are piggy-backed on another message".
> 
> Is this really true? W/o a mU=1 the header can be ignored.  I
> understand the sentiment you're trying to get across (the RMD needs
> to be prepared for piggy-backed headers) but I'm not really sure 
> there's anything normative we need to say anything about it beyond
> what SOAP itself says.  Technically, the sending endpoint could
put 
> any header it wants in a message (not just piggy-backed RM headers)
> and those should be treated no differently.  They may or may
not be 
> processed - but if marked with mU=1 then they must be processed.  I
> think some non-normative text might be more approrpriate - something
> along the lines of: 
>   Since the choice of whether or not to piggy-back RM headers
is 
> made by the endpoint sending the message, in order to ensure optimal
> and successful processing of RM Sequences, endpoints that receive
> RM-related messages should be prepared to process RM Headers that
> may be included in any message it receives. 
> 
> thanks
> -Doug
> __________________________________________________
> STSM | Web Services Architect | IBM Software Group
> (919) 254-6905 | IBM T/L 444-6905 | dug@us.ibm.com
> 
> 

> 
> "Gilbert Pilz" <gpilz@bea.com> 
> 01/05/2007 07:45 PM 
> 
> To
> 
> <ws-rx@lists.oasis-open.org> 
> 
> cc
> 
> Subject
> 
> [ws-rx] Concrete proposal for PR021
> 
> 
> 
> 
> Attached is the concrete proposal for point (2) of PR021. There are
> changes to sections 3.2, 3.8, and 3.9. Most of the changes are 
> editorial; I put all the material related to piggy-backing into a
> common paragraph. The substantive changes consist of: 
> Section 3.2: Additional sentence concerning which
party controls the
> use of piggy-backing. 
> Sections 3.8 and 3.9: Additional statements of
the pattern "The RM 
> [Source | Destination] MUST detect and process any [*] header blocks
> that are piggy-backed on another message". 
> - gp 
> <<...>> <<...>> [attachment
"PR012-case2.pdf" deleted by Doug 
> Davis/Raleigh/IBM] [attachment "PR021-case2-changes.pdf"
deleted by 
> Doug Davis/Raleigh/IBM] [attachment "PR021-case2.pdf" deleted
by 
> Christopher B Ferris/Waltham/IBM] [attachment "PR021-case2-changes.
> pdf" deleted by Christopher B Ferris/Waltham/IBM]