NetNews Usenet Archive 1992 #20/NN_1992_20.iso/spool/comp/object/3474

Text File | 1992-09-08 | 1.8 KB | 44 lines

Newsgroups: comp.object
Path: sparky!uunet!walter!porthos!dancer!haim
From: [e-mail] (24103-kilov)
Subject: Re: Object-Oriented Methodologies - Class Specifications
Organization: Bellcore, Livingston, NJ
Date: Tue, 8 Sep 92 15:58:34 GMT
Message-ID: <[e-mail]>
References: <[e-mail]> <graham.715829976@galois> <[e-mail]>
Sender: [e-mail] (USENET System Software)
Lines: 32

Eric Smith says:

The trick is to see that relationships are themselves objects.  For example,
the ways a line and a circle can relate to each other are the line-circle
relationship, which is a subclass of the two-geometric-figure relationship.
Intersections are a feature of geometric figures, and the intersections of
a line and a circle are a feature of the line-circle-relationship subclass
of the two-geometric-figure relationship.
---

Of course! And, moreover, a relationship is defined exactly like any other
class is defined, i.e., by using available operations. Note that these
operations will be jointly owned by more than one object class (in this
example, by a line and a circle...).

And an operation is defined by appropriate pre- and postconditions that act
within the context specified by an invariant.

This approach is close to the one used in the OSI Managed Relationship
Draft Standard.

We use it for a while in information modeling, and we have a library of
generic relationships used for this purpose (dependencies, compositions,
references, "traditional" (Chen-like) relationships, etc. The framework
of this approach is described in "The Framework" (of course), a Bellcore
publication, and in some papers at conferences like TOOLS '91, TINA '92, etc.

Hope this helps.

-Haim Kilov
[e-mail]      or
[e-mail]