Re: message exchange patterns and # of parties


I feel rather like the high priestess of multicast.  "Is our credo

My feeling when this wording was introduced was that it was adequate.  I
received pushback from some other people (inside and outside the working
group) who did not think so.  At least one person believed that it would
be more correct to differentiate between multicast patterns and unicast
patterns, in the definitions.

Once alerted, I found myself uncomfortable with some of the language,
but the only bit that I can actually identify is the use of the more
restrictive term "broadcast" rather than the more generic "multicast". 
The other problem, that would probably require completely different
wording, would indicate that the *other* receivers of the message *also*
participate, in parallel, in the *same pattern*, which this formulation
implies is *not* the case.  That is, the formulation appears to imply
that each message from the service to another node is to a single other
node; if the message happens to be multicast, those other receivers are
not participants in the pattern.

Language grows awkward here, because the essence of multicast is sending
a single message which has multiple recipients, all of whom may respond
in an identical fashion.

Or, in short: I would prefer to replace "broadcast" with "multicast",
and otherwise would wait for others to argue that this wording is not
okay before trying to do any further editing.

On Thu, 18 Sep 2003 12:33:19 -0400
David Booth <> wrote:

> Amy,
> On today's teleconference I took an action to ping you to see if you
> are satisfied that the wording in the current patterns draft[1] is
> clear enough about permitting multicast.  From my earlier message[2]:
> [[[[
> The patterns document[1] currently says:
> [[
> Like interfaces and operations, WSDL message patterns do not
> exhaustively describe the set of messages exchanged between a service
> and other nodes; by some prior agreement, another node and/or the
> service may send other messages (to each other or to other nodes) that
> are not described by the pattern. For instance, even though a pattern
> may define a single message sent from a service to one other node, the
> Web Service may broadcast that message to other nodes.
> ]]
> That verbiage seems clear enough to me (i.e., it permits the case you
> describe), but maybe that's just because I already know what I think
> it should be saying.  Do you think we need to add more clarification
> to it?  If so, what would you propose?
> ]]]]
> 1.
> 2.
> -- 
> David Booth
> W3C Fellow / Hewlett-Packard
> Telephone: +1.617.253.1273

Amelia A. Lewis
Architect, TIBCO/Extensibility, Inc.

Received on Friday, 19 September 2003 11:11:45 UTC