RE: [wsdm] On ambivalent associations

["Andrea Westerinen" <andreaw@cisco.com>]

Title: Message

The main issue with directionality is that the model semantics do not map easily to it when a direction does not exist.  By using the words, source and target, in the relationship definition, you have explaining to do when they don't apply.  I would prefer to see something like Role1, Role2, Role3, ... (assuming more than binary relationships :-) that supports all the model semantics cleanly, rather than saying "I know that source and target don't apply in this case, but I had to ignore that in order to fit into the WSDM definition."  This makes the current solution wrong, IMHO.

 

Use of terms such as Role1, Role2, Role3 is more inline with base UML definitions of relationships as well.  Directionality and binary relationships are not mandated by UML.

 

Andrea 

-----Original Message-----
From: Sedukhin, Igor S [mailto:Igor.Sedukhin@ca.com]
Sent: Thursday, August 05, 2004 3:51 PM
To: wsdm@lists.oasis-open.org
Subject: [wsdm] On ambivalent associations

We had this discussion today about relationships/associations being directional or not. I'm still trying to understand the reasoning behind this. One may equivalently say

#1 A and B are friends

#2 A is a friend of B and B is a friend of A

I don't understand why #1 is so important.

Handling/modeling #1 gets very intricate when there are more than two participants. "A and B and C are friends". Saying that this has to be represented as "A and B are friends" + "A and C are friends" + "B and C are friends" breaks the pattern here. This will be hard to compute.

At the same time saying that "A is a friend of B" + "A is a friend of C" + "B is a friend of A" + "B is a friend of C" + "C is a friend of A" + "C is a friend of B" does not break the pattern that was already itroduced. This is easy to compute.

The difference is 2x. That is not a significant difference.

-- Igor Sedukhin .. (igor.sedukhin@ca.com)
-- (631) 342-4325 .. 1 CA Plaza, Islandia, NY 11788