Fixing Classification: A Viewpoint-Based Approach

The concept of classification as realized in most traditional object-oriented computer languages has certain limitations that may inhibit its application to modeling more complex phenomena. This is likely to prove problematic as modern software becomes increasingly more integrated with the highly dynamic physical world. In this paper, we first provide a detailed description of these limitations, followed by an outline of a novel approach to classification designed to overcome them. The proposed approach replaces the static multiple-inheritance hierarchy approach found in many object-oriented languages with multiple dynamic class hierarchies each based on different classification criteria. Furthermore, to better deal with ambiguous classification schemes, it supports potentially overlapping class membership within any given scheme. Also included is a brief overview of how this approach could be realized in the design of advanced computer languages.

1. 1.

   https://www.verywellhealth.com/organ-system-1298691.

2. 2.

   https://smartanythingeverywhere.eu/.

3. 3.

   The first European scientists to examine the body of a platypus in 1799 declared it to be a fake, consisting of sewn-together parts of multiple different animals (https://en.wikipedia.org/wiki/Platypus).

4. 4.

   https://en.wikipedia.org/wiki/Tomato#Fruit_versus_vegetable.

5. 5.

   https://en.wikipedia.org/wiki/Digital_twin.

6. 6.

   https://en.wikipedia.org/wiki/Multiple_inheritance.

7. 7.

   This principle is, of course, yet another example of the divide-and-conquer strategy.

8. 8.

   For the purposes of this work, we use the definition of viewpoint as provided in the IEEE 42010 standard [7]. It defines a “view” as a description of a system “from the perspective of specific system concerns”, and a “viewpoint” as the “conventions for the construction and interpretation…of views”.

9. 9.

   Box and Draper in their work on empirical model building, state boldly that “essentially, all models are wrong, (but some are useful)” [1].

10. 10.

    We intentionally chose a different name than “feature” for this real-world element to better distinguish it from its conceptual representation, which is called Feature in this framework.

11. 11.

    The concept of ClassifiableEntity can be viewed as a type in the traditional software interpretation of that term.

12. 12.

    This may not necessarily be an error, but it does imply that the classification schemes are incomplete, which may not have been the intent.

13. 13.

    Smalltalk-80 was a dynamically typed language, so static type checking was impractical in any case.

Authors and Affiliations

1. Malina Software Corp., Ottawa, Canada

   Bran Selić

2. Monash University, Melbourne, Australia

   Bran Selić

3. Università degli Studi dell’Aquila, L’Aquila, Italy

   Alfonso Pierantonio

Corresponding author

Correspondence to Bran Selić .

Editors and Affiliations

1. University of Limerick, Limerick, Ireland

   Tiziana Margaria

2. TU Dortmund, Dortmund, Germany

   Bernhard Steffen

Reprints and permissions

© 2021 Springer Nature Switzerland AG

Policies and ethics