Not sure I follow what you are saying Monica. I am not
saying that we
cannot model it with existing MEP's. Rather I am saying it
is better
to make a notification explicit. Therefore to do this we
need to
distinguish between a response that is paired with a request
as
opposed to a response that is not. The latter would be a
notification. By making it explicit, which does not change
the
semantics nor does it require new MEP's, we simply make it
clear. One
of the profound benefits of WS-CDL descriptions is the
clarity they
bring by describing the interactions. Currently interactions
do not
explicit denote a notification. Rather they enable the
description to
use a response. This is not really correct because it is not
a
response. A response is a response to something. A
notification is
different in terms of how it may be understood. The
semantics (that
that is observable) are identical. So we need some way of
making it
clear that our "response" is not a mistake. Having
an explicit
support for notification would ensure clarity. That is all I
am after
from this.
Cheers
Steve T
On 6 Nov 2006, at 14:26, Monica J. Martin wrote:
>
>> Steve Ross-Talbot wrote: I concur with Charlton. It
does not
>> change the semantics and so has no major impact
but does have a
>> clear benefit of adding clarity to descriptions.
Difficult to see
>> why we should not do it.
>>
>> Cheers
>> Steve T
>
> mm1: In our investigation, we found no substantive
reason not to
> model this with existing MEP already defined and used
in WS-CDL.
> There was no need for a new pattern and in fact it was
redundant.
> Thank you.
>
>
>
|