Real Projects with Informal Models

Dora Dzvonyar, Stephan Krusche, Lukas AlperowitzEdusymp@models 2014

Introduction

In software engineering, the initial phase of a project—often called the "cold start"—is fraught with communication challenges. Developers and customers frequently begin with a significant disconnect: customers may have a vision but lack technical specifics, while developers need concrete requirements to begin building. Traditional formal models, such as complex Unified Modeling Language (UML) diagrams, are often too rigid or intimidating for these early-stage brainstorming sessions. This paper explores how "informal models"—including video trailers, hand-drawn sketches, and "software theater"—can bridge this gap, facilitating a more synchronized and honest exchange of ideas.

Methodology: Navigating the Cold Start

The research is based on a case study of the iOS Praktikum at the Technische Universität München. During this capstone course, 100 students developed 16 mobile applications for 11 industry partners, including major corporations like BMW and Siemens. To manage the diverse starting points of these projects, the authors identified three distinct "Situations":

  1. The Blank Slate (Situation 1): No problem statement or design exists.
  2. The Visionary (Situation 2): A problem statement is provided, but there is no technical architecture.
  3. The Refinement (Situation 3): Both a problem statement and a top-level design are already in place.
Workflow variations based on the initial project state (Situations 1-3).

To address these scenarios, the course utilized "Sprint 0"—a dedicated non-development period focused entirely on alignment and requirements elicitation. During this phase, teams employed several informal techniques:

  • Software Cinema: Creating short video trailers to communicate the "main idea" and visionary scenarios.
  • Software Theater: Acting out user scenarios in real-time during presentations to provide a tangible impression of the app's utility.
  • Informal Top-Level Design: Using icon-based diagrams rather than strict UML to explain system architecture to non-technical stakeholders.

The Strategic Advantage of "Unpolished" Models

A key insight from the study is the "unpolished advantage." Low-fidelity prototypes, such as paper sketches or basic digital mock-ups, often solicit more honest and critical feedback than high-fidelity designs. When a design looks "finished," customers may hesitate to suggest changes, fearing that the effort to modify the work would be too high. Conversely, a rough sketch signals that the design is still fluid and open to significant revision.

The evolution of a user interface from a hand-drawn sketch to a high-fidelity application.

These informal models serve as "social glue," synchronizing technical expertise across a diverse group and building team spirit. The utility of each model is context-driven; for instance, video trailers were essential for discovery in "Situation 1," while informal architectural diagrams were more effective for integration tasks in "Situation 3."

Results and Industry Feedback

The qualitative data from the iOS Praktikum suggests that informal models significantly improve stakeholder engagement. Industry partners reported that early prototypes helped detect misunderstandings during Sprint 0, long before expensive coding began. For example, representatives from BMW noted that these early visualizations were crucial for error detection. Similarly, partners from NTT DATA found that informal, icon-based diagrams conveyed system architecture more effectively to non-technical stakeholders than traditional UML Deployment Diagrams.

The successful delivery of 16 applications demonstrates that starting with informal models does not preclude a professional, high-quality outcome. Instead, it ensures that the final product is more closely aligned with the customer's actual needs.

Conclusion and Future Directions

While informal models are powerful tools for early-stage communication, they are not a total replacement for formal specifications. The paper notes a "transition gap"—the point at which a team must move from informal sketches to formal technical documentation. Future research could focus on creating a more structured framework for this transition. Additionally, while the qualitative feedback is overwhelmingly positive, future studies could benefit from quantitative metrics, such as tracking the number of requirement changes or bug reports in projects using informal versus formal early-stage models. Despite these limitations, the use of "software theater" and low-fidelity prototyping offers a compelling model for managing complex software projects in both educational and industrial settings.

Summary updated

BibTeX
@inproceedings{dzvonyar2014real,
  author = {Dora Dzvonyar and Stephan Krusche and Lukas Alperowitz},
  title = {{Real Projects with Informal Models}},
  booktitle = {Edusymp@models},
  year = {2014},
}
Figure