default fault destination

In some proposals for a transport binding, there's a notion of a
default fault destination; e.g., "when using HTTP, deliver any faults
generated back to the client."

I'm wondering if it's more appropriate to define this in the SOAP
application (e.g. RPC), and leave the transport agnostic about fault
delivery.

Thoughts?


-- 
Mark Nottingham
http://www.mnot.net/

Received on Friday, 20 July 2001 18:16:06 UTC