Browse Prior Art Database

Graphical Object Representation

IP.com Disclosure Number: IPCOM000120600D
Original Publication Date: 1991-May-01
Included in the Prior Art Database: 2005-Apr-02
Document File: 2 page(s) / 68K

Publishing Venue

IBM

Related People

Cavendish, C: AUTHOR [+2]

Abstract

Object-oriented systems are being developed with a primary goal of generating reusable code components called classes. The difficulty of managing the library of classes of objects can quickly escalate as a system develops and the number of classes in the library grows larger. Current object-oriented systems sometimes provide a class browser which lists all the available classes in a list form and may also provide a graph of the class hierarchy. However, this is inadequate if the number of classes is large. A new method is provided for graphically representing and browsing the class library which reduces the complexity of developing systems with large numbers of classes.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

Graphical Object Representation

      Object-oriented systems are being developed with a
primary goal of generating reusable code components called classes.
The difficulty of managing the library of classes of objects can
quickly escalate as a system develops and the number of classes in
the library grows larger.  Current object-oriented systems sometimes
provide a class browser which lists all the available classes in a
list form and may also provide a graph of the class hierarchy.
However, this is inadequate if the number of classes is large.  A new
method is provided for graphically representing and browsing the
class library which reduces the complexity of developing systems with
large numbers of classes.

      The technique provided graphically represents classes as nodes
of a series of undirected graphs.  Some of the graphs are acyclic and
some are cyclic.  The edges of the graph have an associated distance
measure which represents the closeness or proximity of one class to
another.  The distance values for the edges will be small for objects
which are similar and large for objects which are dissimilar.  An
algorithm is provided for deriving the distance measure from the
definition of the classes.  A dictionary of methods will provide a
way to compare the similarity of classes.

      For example, if object A and object B both perform 'open',
'close', 'maximize' and 'minimize' they would probably have low
proximity numbers.  However, if object C did not have any methods
similar to object A, the edge from A to C would be v...