Investigating Differences between Graphical and Textual Declarative Process Models

TL;DR

This study compares graphical and textual declarative process models, finding that graphical models are easier to understand despite potential semantic confusions, while textual models increase error rates and cognitive load.

Contribution

It provides empirical evidence on the cognitive differences between graphical and textual declarative process modeling not previously quantified.

Findings

Graphical models are easier to understand than textual models.

Textual models lead to higher error rates and mental effort.

Graphical lookalikes cause confusion in declarative modeling.

Abstract

Declarative approaches to business process modeling are regarded as well suited for highly volatile environments, as they enable a high degree of flexibility. However, problems in understanding declarative process models often impede their adoption. Particularly, a study revealed that aspects that are present in both imperative and declarative process modeling languages at a graphical level-while having different semantics-cause considerable troubles. In this work we investigate whether a notation that does not contain graphical lookalikes, i.e., a textual notation, can help to avoid this problem. Even though a textual representation does not suffer from lookalikes, in our empirical study it performed worse in terms of error rate, duration and mental effort, as the textual representation forces the reader to mentally merge the textual information. Likewise, subjects themselves expressed that the graphical representation is easier to understand.