W3C home > Mailing lists > Public > www-ws-arch@w3.org > June 2003

RE: SOAP UML diagram

From: Martin Chapman <martin.chapman@oracle.com>
Date: Thu, 12 Jun 2003 12:58:36 -0700
To: "Cutler, Roger \(RogerCutler\)" <RogerCutler@chevrontexaco.com>, <www-ws-arch@w3.org>
Message-ID: <PEEBJKKCFNCENDPJDEMIAECDDFAA.martin.chapman@oracle.com>



> -----Original Message-----
> From: Cutler, Roger (RogerCutler) [mailto:RogerCutler@chevrontexaco.com]
> Sent: Thursday, June 12, 2003 12:10 PM
> To: Martin Chapman; www-ws-arch@w3.org
> Subject: RE: SOAP UML diagram
>
>
> Some questions, mostly but not all about cardinality:
>
> A Message can have 0 senders and/or 0 receivers (although the underlying
> thing has 1 on bothe).  Is this right?  If a message is anonymous does
> it have zero senders or is the sender "anonymous"?

not sure what you mean by anonymous here. This goes back to the point of
whether
there has to be at least one sender and at least one receiver for a message.
If we want
to capture the unreliable nature, then I think they should stay at *. But
I'm open to discussing this assumption.

>
> Can you have a Protocol Binding without an Underlying Protocol?

No - the protocol binding is modelled as an association class i.e. it's a
set
of properties related to the association between node and underlying
protocol.
Note you can have an underlying protocol without a protocol binding, but
that doesn't mean much!

>
> Can you have a sender without a message?  (Is that saying that the role
> does not have to be exercised?)

not really, sender is a role wrt to a particular message.
>
> I'm confused about what a Module is.  Why must it have at least one
> Header Block, but a Feature can exist without a Header Block?  Is a
> Module a necessary concept here?

These are all taken from the document section 3.
(http://www.w3.org/TR/2003/PR-soap12-part1-20030507/#extensibility)
Features are a general concept and they may or may not be related to modules
and headers.
Modules describe the semantics of one or more headers e.g a reliability
module.
One ambiguity to me is if a feature may refer to a header without reference
to a module (I believe the spec doesn't prohibit this)?


>
> You've got 0..1 for Header inside Envelope.  I thought a header was
> mandatory.

No its optional, section 5.1:
   ".1 SOAP Envelope
	The SOAP Envelope element information item has:

	A [local name] of Envelope .

	A [namespace name] of "http://www.w3.org/2003/05/soap-envelope".

	Zero or more namespace qualified attribute information items amongst its
[attributes] property.

	One or two element information items in its [children] property in order as
follows:

		- An optional Header element information item (see 5.2 SOAP Header).

		- mandatory Body element information item (see 5.3 SOAP Body)."


>
> Shouldn't MEP have connections to a bunch of other things?  Like Sender,
> Receiver and Node?
>It seems to me that if a MEP is an abstract
> definition of how a bunch of messages are supposed to work together,
> that the various pieces of that pattern need to know that they are part
> of that MEP, don't they?

This is captured implicitly.  An Mep is_a feature (triangle between feature
and MEP), and a feature may constraint the behaviour of one or more nodes.
If you think we need something more explicit I can put an association
between mep and node, with 1..*  on the node end (or even 2..*?).



> -----Original Message-----
> From: Martin Chapman [mailto:martin.chapman@oracle.com]
> Sent: Thursday, June 12, 2003 12:24 PM
> To: www-ws-arch@w3.org
> Subject: RE: SOAP UML diagram
>
>
>
> updated diagram at:
>
> http://lists.w3.org/Archives/Public/www-archive/2003Jun/0019.html
>
>
>
>
Received on Thursday, 12 June 2003 15:58:29 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 3 July 2007 12:25:21 GMT