Business process management is a multi-billion-dollar industry focused on modeling business processes to analyze, improve, and automate them. These processes are interconnected activities an organization follows to achieve its goals. Modeling is essential to understand, manage, and boost organizational efficiency and competitiveness.

Characterizing modeling artifacts is essential in empirical software engineering, with prior studies exploring domains like machine learning, automotive, video games, and bots. Given the rise in organizational use of process models, we focus on business process models. To our knowledge, no characterization of open-source business process models exists in the literature, which we aim to address by answering the following research question:

RQ1: What is the landscape of business process models in open source?

By characterizing business process models in open-source environments, we aim to inform (a) the development of tools suited to the unique needs of open-source business process management projects, (b) influence best practices for model creation and maintenance, and (c) contribute insights for refining business process modeling standardization efforts and future revisions.

Our study shows that business process models on GitHub span at least 16 domains, with machine learning, traditional software, sales, business services, and financial services being the most common. Business process model adoption is rising, reflected by the steady increase in model creation and updates. While organization-owned repositories contribute significantly (approximately 16%), most models (approximately 84%) are hosted in individual-owned repositories. Additionally, we find a strong reliance on tools from industry-leading vendors like Drools, Camunda, Activiti, and SAP Signavio, underscoring their importance in the open-source business process modeling community.

Our characterization explores both the diversity and reuse of business process models, emphasizing cloning as a critical factor impacting model quality. Initial exploration of GitHub models reveals significant duplication, which can complicate maintenance, evolution, and quality assurance. Although clone detection is well-studied in software engineering, especially for source code, an analysis of business process model similarity in open-source settings is still missing. Thus, we expand our business process model characterization by addressing our second research question:

RQ2: What is the extent of cloning in business process models in open source?

We find that the majority of clones (approximately 80%) in business process models were found to be exact clones, with cross-project clones being the most common (approximately 57%) and predominantly associated with industry-owned (as opposed to academia-owned) repositories. An analysis of the model-level clones indicates that machine learning, traditional software, and business services are the top three domains in the dataset. A further clone detection analysis on subprocess elements within the models reveals a substantial degree (approximately 92%) of cloning among these model fragments. It also shows that all the detected subprocess clones have at least one exact clone.

In summary, the primary contributions of this study are:

First study characterizing open-source business process models on GitHub. First study on business process model and subprocess-level cloning on GitHub. A unique dataset with tagged, distinct, non-trivial business process models, their similarities, and distinct subprocesses.