From: George Beeler [woody@beelers.com]
Sent: Thursday, January 30, 2003 8:17 PM
To: 'HL7-CQ'
Subject: FW: Modeling of Message-to-Acknowledgement associations

Since Lloyd's note (below) makes complete sense to me, (with the change to allow a Message to be the subject of many acknowledgements), I propose to make this as a change to the next RIM, in order that good message design can be done by CQ.  If anyone in CQ objects, "speak now or forever...."

Yours .... Woody Beeler

> -----Original Message-----
> From: Lloyd McKenzie [mailto:lmckenzi@ca.ibm.com]
> Sent: Thursday, January 30, 2003 10:53 AM
> To: George Beeler
> Cc: 'HL7 CQ (E-mail)'; Dale@ZED-logic.com
> Subject: Re: Modeling of Message-to-Acknowledgement associations
> 
> 
> Hi Woody,
> 
> Rather than answer your questions, I'll propose what I think
> we need, including improved labels, definitions, and 
> reversing the cardinalities.
> 
> What we need is the following:
> 
> Message: conveysAcknowledgment (0..1) Acknowledgment :
> conveyedBy (1..1)
> Definition: Identifies the relationship between an 
> acknowledgment and the message conveying that acknowledgement.
> 
> Message: acknowledgedBy (0..1) Acknowledgement : acknowledges (1..1)
> Definition: Identifies the relationship between a message and
> the acknowledgement that acknowledges that message.
> 
> AcknowledgmentMessage: acknowledgment (1..1) Acknowledgment :
> annotatedBy
> (0..*)
> Definition: Identifies the relationship between an 
> acknowledgment and the error, warning and informational 
> messages accompanying that acknowledgment.
> 
> 
> Lloyd
> 
> Lloyd McKenzie, P.Eng.              I/T Architect, IBM Global Services
> Internet: LMCKENZI@CA.IBM.COM
> PhoneMail: (780)421-5620          Internal Mail:04*K0R*1004 *EDM
> 
> 
> "George Beeler" <woody@beelers.com>@lists.hl7.org on
> 2003-01-29 07:41:21 PM
> 
> Please respond to "George Beeler" <woody@beelers.com>
> 
> Sent by:    owner-cq@lists.hl7.org
> 
> 
> To:    "'HL7 CQ \(E-mail\)'" <cq@lists.hl7.org>
> cc:    <Dale@ZED-logic.com>
> Subject:    Modeling of Message-to-Acknowledgement associations
> 
> 
> 
> All -
> 
> Dale Nelson raised a Visio question that got me to focus on the set of
> classes: Message, Acknowledgement, and
> AcknowledgementMessage.  This, in turn has raised questions 
> about the correctness of the model.  I have attached two 
> gif's as illustration.
> 
> The first, MessageAcknowledgementModel is an introduction to
> the topic that shows the three classes as they currently 
> exist in RIM 1.20.  I included the verbatim descriptions of 
> the three classes, and of the two associations that have descriptions.
> 
> The second drawing, MessageAcknowledgementQuestion, includes
> overlays expressing my understanding of the associations, and 
> highlights the questions I have.  In order:
> 
> A) Moving down the right-hand association from Message to
> Acknowledgement, I believe you are moving from the header of 
> an acknowledgement message down to the content of the acknowledgement.
> 
> 
> B) Moving from Acknowledgement down to
> AcknowledgementMessage, I believe you are moving to a 
> particular annotation in the acknowledgement. (At first I 
> thought this was a stub that might connect to a message body 
> in the way that a body is linked to ControlAct.  If this is 
> intended, none of the documentation gives a clue that this is so.)
> 
> C) Moving up the left-hand association from Acknowledgement
> to Message, I believe you are walking the link from the 
> Acknowledgement to the "header" of the message being acknowledged.
> 
> Pre-question -- Is this basic understanding correct?  If it
> is not, then my following "real" questions are founded on a 
> misunderstanding.  If my understanding is correct, I submit 
> that someone should revise the descriptions of the three 
> associations to say essentially what I have said in the 
> diagram annotations.
> 
> Question 1: Indicated by the red circles and numeral one.
> Are the cardinalities reversed on the association from 
> Acknowledgement to AcknowledgementMessage?  From the 
> description, I can imagine a given acknowledgement having 
> zero to many annotations.  Moreover, I cannot imagine an 
> annotation referring to more than a single acknowledgement.
> 
> Question 2: Should BOTH of the cardinalities (circled in
> orange) from Acknowledgement to Message be mandatory (1..1)?  
> The implication is that whenever an Acknowledgement is used 
> in an HMD, it should link to both its own header and the 
> header of the message it is acknowledging.  To be honest, I 
> have not found good use case for leaving either one out.  
> Therefore, I suspect these cardinalities are correct.
> 
> 
> If I can get a quick response to these questions, I will
> include the changes in RIM 1.21 which will come out "soon."
> 
> Yours .... Woody Beeler
> 
> George W. Beeler, Jr. Ph.D.
> Beeler Consulting LLC
> 807 Tenth Ave. SW
> Rochester, MN 55902-6357  USA
> 
> email: woody@beelers.com
> voice: 507-254-4810 (& voice mail)
> fax: 507-281-8951
> 
> 
> 
> 
> 
> 
> 
> 
