Re: another approach to status codes, etc. in HTTP binding

7/19/2001 2:29:26 PM, Rich Salz <rsalz@zolera.com> wrote:

>> The way I'm thinking presently is that a we provide that a SOAPFault
>> MAY be carried on an HTTP 500, but this is NOT a requirement
>> as is currently specified in 1.1 and 1.2WD.
>
>Horrible.  Let's pick one place and tell folks to stick it there.  If we
>have to count angels on pins first, so be it.

Yep.  I could *almost* see saying both (200 if tunnelling, 400/500 if not), except that the typical developer won't know 
the difference, so that wouldn't be very helpful for interoperability.

What about two bindings?  One for respecting HTTP's semantics, and the other for abusing them. 8-)

MB