Re: [ws-caf] Issue: mustUnderstand values

["Mark Little" <mark.little@arjuna.com>]

Title: Message

I think we need another issue, because I don't remember this coming up at New Orleans.

 

Mark.

 

----
Mark Little,
Chief Architect, Transactions,
Arjuna Technologies Ltd.
 
www.arjuna.com

----- Original Message -----

From: Furniss, Peter

To: Mark Little ; Guy Pardon ; ws-caf@lists.oasis-open.org

Sent: Wednesday, May 26, 2004 11:36 AM

Subject: RE: [ws-caf] Issue: mustUnderstand values

We seem to agree.

 

Draft 0.2, 2.1 says "Hence in this case [context with application message] each SOAP header block carrying a context has the
“mustUnderstand” attribute set to “true” and the recipient must understand the header block
encoding according to its identifying URI."

 

Is this changing arising from New Orleans or do we need an issue ?

 

Peter

-----Original Message-----
From: Mark Little [mailto:mark.little@arjuna.com]
Sent: 26 May 2004 10:49
To: Furniss, Peter; Guy Pardon; ws-caf@lists.oasis-open.org
Subject: Re: [ws-caf] Issue: mustUnderstand values

I think mustUnderstand may be set by referencing specifications, but there's nothing we should do in Context.

 

Mark.

 

----
Mark Little,
Chief Architect, Transactions,
Arjuna Technologies Ltd.
 
www.arjuna.com

----- Original Message -----

From: Furniss, Peter

To: Guy Pardon ; ws-caf@lists.oasis-open.org

Sent: Tuesday, May 25, 2004 6:41 PM

Subject: RE: [ws-caf] Issue: mustUnderstand values

Actually, do we need to mandate a value for the SOAP mustUnderstand attribute ? It would seem plausible for a context to be sent "on spec" - to be understood if the receiver understands it, but can be ignored if it doesn't and no harm will be done.

 

The equivalent construct caused a long discussion in OTS circles a few years ago - was it wrong to send transaction context to a service that didn't understand it. In this case, to assert that it is always wrong to send ANY kind of context to a service that might not understand it seems to be going too far.

 

Leaving mustUnderstand to the sender (or possibly the referencing specification) would seem to allow exactly the right flexibility - if the sender's anticipations will be compromised by non-support of the context (of the specific type of context ??), then it can put mustUnderstand=1. If the context effect is "do this/use this if it applies to you, otherwise don't worry", then it can put mustUnderstand=0.

 

 

 

Peter

-----Original Message-----
From: Guy Pardon [mailto:guy@atomikos.com]
Sent: 25 May 2004 16:47
To: ws-caf@lists.oasis-open.org
Subject: [ws-caf] Issue: mustUnderstand values

Hi,

This should actually go into the issues system (I will post it there as soon as I can):

AFAIK, the mustUnderstand = "true" attribute value doesn't seem to be WS-I compliant (should use 0/1, not false/true).

The CTX spec draft 0.2 mentions this in several places...

Guy

Dr. Guy Pardon ( guy@atomikos.com )

Atomikos: Your Partner for Reliable eBusiness Coordination

http://www.atomikos.com/

The information in this email is confidential and only meant for the addressee(s). The content of this email is informal and will not be legally binding for Atomikos.