Sometimes it is said that excellent software artifacts are also aesthetically appealing, representing some sense of beauty for an observer. Thriving systems theory (TST), an information systems (IS) design theory, sets out to operationalize aesthetic quality attributes of information systems, such as elegance and symmetry.

The paper is based on the often-cited work of Christopher Alexander in design theory for physical architectures (such as buildings). The authors translate Alexander’s 15 properties of centers of living structures into the IS domain. In Alexander’s conception, these center properties (such as local symmetries, levels of scale, echoes, simplicity, and inner calm) serve both as a means to consistently assess the degree of order of an artifact, and as elements to be manipulated when designing the artifact.

The adage that “beauty lies in the eye of the beholder” is deconstructed into a four-part experience: (1) the observer’s mindset; (2) the observer’s expectation; (3) a threshold, that is, the encounter between the system and the observer; and (4) an underlying system realization.

This approach renders TST prescriptive, in the sense that design achieves the alignment of an architect’s choices with the aggregate intentions of the stakeholders, by informing observer expectations, adjusting individual thresholds (interfaces), or adjusting choices in the 15 center properties.

Eventually, this yields--on a fairly superficial level--a not uninteresting amalgam of choice properties well known in current software architecture, such as modularization, cohesion, and encapsulation. There are also some more esoteric properties, such as identity and elegance. Unfortunately, readers will have to turn to several other papers by the authors to find successful applications of TST for assessing design quality in modeling.

Personally, I very much subscribe to the idea that “beautiful equals high quality” in software code. However, I find the paper not entirely convincing, because the theory is largely operating in the visual domain, whereas both software and physical artifacts also exhibit properties through usage, and those properties (also known as nonfunctional requirements such as performance) also determine quality for stakeholders. Nevertheless, this easily accessible paper provides a nice digression from mainstream metrics-based design theory, and maybe TST will evolve into a more exhaustive design theory in the future.