Re: updated draft to put F&P in more places

Hugo Haas wrote:
> * Sanjiva Weerawarana <sanjiva@watson.ibm.com> [2004-06-25 23:29+0600]
> 
>>I've updated the draft per the telecon decision yesterday to put
>>F&P in a few more places. Please review.
>>
>>http://dev.w3.org/cvsweb/~checkout~/2002/ws/desc/wsdl20/wsdl20.html
>>
>>I note (with disgust) that we're severly discriminating against
>>Interface Fault components and Binding Fault components by not 
>>allowing them to contain F&P (properties). Poor suckers.
> 
> 
> I actually don't understand why this is either. Could somebody explain
> the issue around having F&P on fault components?

There are no issues. Perhaps the only puzzling aspect of adding them
that I can see is that in interfaces there'd be two places to put
f&p's that relate to a given fault: the interface fault component and
all the fault reference components that use it. Of course f&p's
on the interface fault component apply across all operations, whereas
those on fault reference components only apply to the containing
operation.

By the way, in the spirit of hoisting faults to the same level as
operations, I think that if we had to choose between interface faults
or fault references as the only place f&p's are allowed to appear at,
I'd choose the former.

So +1 to fixing this anomaly.

Roberto

Received on Monday, 28 June 2004 09:47:15 UTC