RE: Issue 146

Hi Doug,

I think maybe we have our wires crossed a bit here.

> -----Original Message-----
> From: Doug Davis []
> Sent: 05 November 2001 15:15
> To:
> Subject: Re: Issue 146
> Stuart Wrote:
> >"The terms 'default actor', 'anonymous actor', 'ultimate recipient' and
> >'SOAP Endpoint' are being used loosely as synonyms. It is not clear
> >a SOAP Node acting as default actor, anon actor, ultimate recipient, or
> >endpoint with respect to a given SOAP message may behave as SOAP
> >intermediary and relay the SOAP message to further SOAP Nodes. Stated
> >differently: Does a SOAP Message Path always terminate at the default
> >anonymous actor, ultimate recipient or SOAP endpoint?"
> >Before, substantively discussing the issue, I would first ask for
> >confirmation that the above sufficiently describes the issue.
> Reading the issue 140 (How does an actor know that it is the ultimate
> recipient?), I don't think the questions you asked not are the right ones
> at least not the way I read it.

Issue 140 is closed it's short description is "How does a SOAP node
determine if it is the anonymous actor?". The discussion of 140 broadened to
be "How does a SOAP node determine if it is a given actor with respect to a
particular message?"

The closure of 140 IMO covers that broader question. However, the discussion
on the thread about 140 continued beyond the resolution of 140 to question
whether or not the anonymous actor could relay a message further along a
message path. Take a look at one of your own messages [1], I don't think the
answer you give there is as clear cut as the emphatic "yes" you give below.

This Issue became numbered Issue 146, and the description of 146 became "How
does an actor know that it is the ultimate recipient?" which is barely if at
all distinguishable from the the original 140. When we came to discuss 146
on the WG telcon, it did not look like issue 146 was properly described in
the issues list. I was then tasked with trying get a clear articulation of
issue 146.

The the issue 140 thread forks around the message at [2] and what I am
trying to do is briefly describe what the issue is that is being discussed
in [2] and the remainder of that thread.

> I interpreted the issue more a long
> the lines of "how does a SOAP node know whether or not it is acting
> as an intermediary or as the ultimate recipient?"

The discussion in [1,2] and beyond seems to challenge whether or not the
ANON actor can act as an intermediary.

> The questions you've asked are looking at a different issue - you're
> focused on whether a message path ends at the ultimate recipient - and
> I believe the answer to that one must be "yes" (in the non-error case).

Yes... but we also seem to use default and anon to denote the ultimate
recipient, and at least in [2] you don't seem so comfortable with anon being
the ultimate recipient.

> If the ultimate recipient does relay the SOAP message on to a new SOAP
> node then that doesn't change the role(s) that Node was playing.  It is
> perfectly valid for a SOAP node to initiate a "new" message path - which
> is what I believe your questions are eluding to - the key being that any
> forwarding from the ultimate recipient is a "new" message and a "new"
> message path.

Yes... I'd agree with that.

> To me, the question  is whether that "forwarding"
> that is taking place is because a new message path is being initiated,
> or whether it is taking place because this node believes it is not the
> ultimate recipient.  Right now the spec does not indicate how a Node
> should make this determination - nor should it (IMO).
> To me, the more generic question should be - "how does a SOAP node know
> roles it is playing?"  

That feels like 140 again... and i think that got adequately closed (at
least for me).

> Whether that determination is getting the list
> of actors it plays, or whether it is determining whether or not it is
> an intermediary or the ultimate recipient - it's all related.  And all
> outside the scope of the spec - it's an exercise left up to the
> coder.  8-)

I think thats how 140 got closed.

> -Dug



Received on Monday, 5 November 2001 12:01:42 UTC