Using the “Physics” of Notations to Analyze a Visual Representation of Business Decision Modeling John C. Thomas, Judah Diament, Jacquelyn Martino, Rachel K. E. Bellamy IBM T. J. Watson Research Center Yorktown Heights, New York USA {jcthomas, djudah, jmartino, rachel}@us.ibm.com Abstract—Visual representations are common in communicating about artifacts such as computer programs, software architectures, and business rules. Yet, generally speaking, these representations seem much harder to learn and to use than many of the representations in other domains. Moody [1] has pointed this out and proposed a set of principles for visual representations based on a wide review of relevant literature in cognitive psychology and software engineering. The real test of this framework is to use it. In this paper, we apply the principles set forth in Moody to examine and improve a proposed representation for business rules and business decisions.. Keywords-business rules, business decisions, decision tables, visual representations, human-computer interaction I. INTRODUCTION Addressing complexity issues, both theoretical and practical, concerns several fields [2]. One important domain, engineering complex software systems, relates to creation, modification and comprehension via visual abstractions. In “Physics” of Notations, Moody [1] argues that visual representation methods in software engineering frequently mismatch what is known about human visual processing. Business decision modeling shares many characteristics with visualizing software systems. In particular, business users who are not trained in formal logic need tools to create, modify and understand formal rules in order to run their business [3]. Our research focuses on providing business users methods to represent formal logic. We argue that Moody’s analytic framework bootstraps an improvement to our notation system for business decision modeling. An improved system would minimally allow users to 1) validate business operations as complete and consistent 2) document compliance with laws and regulations as well as with internal policies 3) provide work aids and training materials inputs for employees 4) provide requirements engineering basis for automation. The uptake of a business-users-centric notation for modeling decisions will depend heavily on the representation’s usability and understandability. Iterative development with real users is the “Gold Standard” to predetermine usefulness and usability [4]. Cognitive modeling [5], ethnographic research [6], and heuristic evaluation [7] are also useful techniques. Generally, each of these methods relies on the ability to specify a particular set of users, tasks, and contexts of use. Evaluating a general representation is challenging since a variety of users, tasks, and contexts can conjointly determine its overall utility and usefulness. Conducting one of the above evaluation methods can become elusive and subjective. Since the “determination” of a system as usable built on one set of assumptions may differ from the actual usage in terms of users, tasks, or contexts. We argue that application of an analytic framework is more sensible and less likely to result in surprises of the type catalogued in Thomas & Kellogg [8]. Various methods of examining usability generally, or visual languages, in particular, need not be mutually exclusive alternatives. An analytic method such as Moody’s may also complement other methods. For example, heuristic evaluation, usability tests, and cognitive modeling excel at uncovering problems, but in and of themselves, do not typically reveal the solutions. An analytic framework can suggest why something is problematic for users and thereby offer seeds of a solution. A number of analytic frameworks exist in human computer interaction; e.g.[9][10]. Such techniques seem especially applicable when discussing a general representation for widespread use, such as ours. Specifically, while our target is the business user, other user types, contexts and tasks include: computer programmers who implement rules, business analysts who understand and rationalize the business and encode rules, auditors and quality assurance personnel who read rules to ensure compliance, and executives who evaluate whether the current rule set is actually good for the business. II. ANALYZING AND MODIFYING A NOTATION BASED ON MOODY’S PRINCIPLES Moody presents nine principles for designing effective visual notations: Semiotic Clarity, Perceptual Discriminability, Semantic Transparency, Complexity Management, Cognitive Integration, Visual Expressiveness, Dual Coding, Graphic Economy, and Cognitive Fit. In this section, for seven of the nine principles, we describe how we applied each as we analyzed our current decision management notation. The remaining two principles: Cognitive Fit and Complexity Management we leave aside. Cognitive fit concerns the use of different visual dialects for different tasks and audiences. Our preliminary analysis does not yet extend beyond the one representation. Complexity management states that explicit mechanisms should be used to deal with complexity. In dealing with complexity, it seems important to consider together both the visual representation and how the users interact with that visual representation. We have yet to specify all the interaction mechanisms for our system so we are still in the process of dealing with Complexity Management. One of the mechanisms already employed statically is the decision table, which collapses many related rules into one easily accessible location. 2012 IEEE Symposium on Visual Languages and Human-Centric Computing 978-1-4673-0853-3/12/$31.00 c 2012 IEEE 41