Formal approach to modeling of modern Information Systems

Bálint Molnár, András A. Benczúr, András BéleczkiJournal of the Association for Information Systems 2016 · DOI

Abstract

Most recently, the concept of business documents has started to play a double role. On one hand, a business document (word processing text or calculation sheet) can be used as a specification tool; on the other hand, the business document is an immanent constituent of business processes, and thereby an essential component of modern Information Systems (IS). The recent tendency is that the majority of documents and their contents within business IS remain in semi-structured formats, while only a smaller portion is transformed into structured database schemas. In order to manage this complexity, we suggest a theoretical framework for modeling business IS and a design method based on a theoretical model that provides structuring principles. This approach focuses on documents and their interrelationships with business processes to assist in perceiving the activities of modern IS.

Introduction: The Shift to Document-Centric Systems

Modern Information Systems are undergoing a paradigm shift. We are moving away from rigid, structured relational databases toward a reality where semi-structured data and web services dominate. In the era of REST APIs and E-government, the "document" has taken center stage. Documents—such as XML/JSON files, spreadsheets, or digital forms—now play a dual role: they serve as both the specification of a process and the actual data flowing through that process.

The Modeling Challenge: Semi-Structured Data

Traditional modeling tools, such as standard Entity-Relationship (ER) diagrams, often struggle to capture the nuances of modern systems. They are particularly ill-equipped to handle:

  1. Semi-structured data: Information that possesses an internal hierarchy but does not fit neatly into flat tables.
  2. The Document Life Cycle: The evolution of a document from a template (with empty fields) to a "ground document" (finalized data).
  3. Complex Interdependencies: The intricate web of relationships between organizational roles, business activities, and the data objects they manipulate.
Hierarchical structure of document objects and their internal elements, illustrating the complexity of semi-structured data.

Methodology: Hypergraphs and Description Logics

To address these challenges, we propose a formal approach based on two mathematical pillars:

  • Generalized Directed Hypergraphs: Unlike standard graphs where edges connect two nodes, hyperedges can connect any number of vertices. This allows for the representation of n-ary relationships, such as a single business process involving multiple documents and roles simultaneously. This recursive structure mirrors the nested nature of modern documents.
  • Description Logics (DL): To ensure the model is computable, we employ Description Logics. This provides a formal language to define "bindings"—the transition of a "free variable" (an empty field) into a "bound variable" (specific data).

This framework integrates these tools with Enterprise Architecture standards like TOGAF and the Zachman Framework, ensuring the model remains aligned with high-level business planning.

Key Insights: The "Crust" and Document States

We visualize the Information System as being surrounded by a "crust" of documents. Whether data enters from a user or is exchanged between services, it must pass through this document layer. A critical distinction is made between intensional documents (rules and templates) and extensional documents (actual data instances). The system's primary function is to transition documents from "free" states to "ground" states.

Key Results: Formalizing the Life Cycle

The primary output of this research is a qualitative framework that provides a formal vocabulary for "Architecture Describing Hypergraphs." This accounts for vertices (data), arcs (processes), and hyperedges (complex relationships).

A formal representation of document states and transitions within a business process life cycle.

By mapping organizational structures directly onto data structures through homomorphisms, we can visualize how documents evolve through business processes. This life cycle model allows for the formal verification of system integrity as data moves through various action states.

Conclusion and Future Work

While the hypergraph approach is mathematically robust, we acknowledge the need for schemes that balance computability with human usability. Future work will focus on developing specialized design tools to assist in the validation of these models and implementing the framework within large-scale E-government case studies to test practical overhead.

Summary updated

Abstract

Most recently, the concept of business documents has started to play double role. On one hand, a business document (word processing text or calculation sheet) can be used as specification tool, on the other hand the business document is an immanent constituent of business processes, thereby essential component of business Information Systems. The recent tendency is that the majority of documents and their contents within business Information Systems remain in semi-structured format and a lesser part of documents is transformed into schemas of structured databases. In order to keep the emerging situation in hand, we suggest the creation (1) a theoretical framework for modeling business Information Systems; (2) and a design method for practical application based on the theoretical model that provides the structuring principles. The modeling approach that focuses on documents and their interrelationships with business processes assists in perceiving the activities of modern Information Systems.
BibTeX
@article{molnar2016formal,
  author = {Bálint Molnár and András A. Benczúr and András Béleczki},
  title = {{Formal approach to modeling of modern Information Systems}},
  journal = {Journal of the Association for Information Systems},
  year = {2016},
  doi = {10.12821/ijispm4404},
}
Figure