RE: An analysis of mustUnderstand and related issues

ok, let me try one more time (more directly).  I believe the
current spec is broken.  If the consensus of the group is
that in the current SOAP spec says that targeted MU headers
reaching the ultimate destination should be ignored, then
I believe that we have a useless feature.  No client should
ever target an MU header because they can not reliably
count on it being delivered to the desired actor, nor will
they be given any indication that those headers are ignored.
Therefore no one should (or will) use it - so I'd like to
change this useless feature into a positive one by changing
the spec to indicate that they should not be ignored.

-Dug


"Henrik Frystyk Nielsen" <henrikn@microsoft.com>@w3.org on 05/16/2001
12:42:03 PM

Sent by:  xml-dist-app-request@w3.org


To:   Doug Davis/Raleigh/IBM@IBMUS, <Noah_Mendelsohn@lotus.com>
cc:   <xml-dist-app@w3.org>
Subject:  RE: An analysis of mustUnderstand and related issues




Isn't this the case with all new features? If you don't use it you don't
use it. However, as one of the fundamental pieces of SOAP is the
extensibility mechanism adding stuff should not have to impact the core
protocol.

One of the points that Noah brings up is if the extensibility mechanism
indeed is good enough and that is a valid concern. Note, however, that
SOAP is quite clear on that it does not define any ordering mechanism
and it is not at all clear to me that it has to as part of the core
spec.

>Perhaps I wasn't clear - if someone doesn't use the feature
>then they're stuck using the current SOAP spec, and if we
>don't change the current spec then they're also stuck with the
>current ambiguity. So, I believe that whether or not we do a
>feature like this we still need to update the spec. -Dug

Henrik Frystyk Nielsen
mailto:henrikn@microsoft.com

Received on Wednesday, 16 May 2001 14:22:30 UTC