W3C home > Mailing lists > Public > www-ws-desc@w3.org > October 2003

Re: writing up faults and q about fault rules

From: Amelia A. Lewis <alewis@tibco.com>
Date: Mon, 27 Oct 2003 09:46:47 -0500
To: Sanjiva Weerawarana <sanjiva@watson.ibm.com>
Cc: www-ws-desc@w3.org
Message-id: <20031027094647.20bd4173.alewis@tibco.com>

On Fri, 24 Oct 2003 15:19:30 +0600
Sanjiva Weerawarana <sanjiva@watson.ibm.com> wrote:
> I'm writing the fault stuff and had a question. The syntax is:
> <(in|out)fault messageReference="ncname" details="qname"/>
> The question is: for an operation that uses a MEP which employs
> message-triggers-fault, what is the relationship between the fault
> direction ("in"/"out") with the direction of the referred to
> message? It seems to me that if message-triggers-fault is used,
> then the if the message is going from service to client (i.e.,
> an "out" message), then the fault will travel in the opposite
> direction, and vice-versa. 

That would be my interpretation as well.

> Is that correct? If so, then how should that fault be declared?
> That is, if its an out-message, should the fault say <infault> 
> and vice-versa? It seems a bit counter-intuitive, but that seems
> to be the right thing if we want to retain infault/outfault.

*laugh*  I'm not fond of outfault/infault, but I believe I'm outvoted. 
Your summary looks completely correct to me.

> Note that for fault-replaces-message, the fault direction and
> message direction MUST always be the same.
> Did I mention that I favor just using "fault" and that I've
> been campaigning to drop the virtual {direction} property of
> message and fault reference components? ;-)

+1 (can my multiple personalities vote?  +asManyAsWeAreWorth, in that
case ...)

(all of us)
Amelia A. Lewis
Architect, TIBCO/Extensibility, Inc.
Received on Monday, 27 October 2003 09:46:42 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:54:45 UTC