Project Management for the 21st Century: Bridging HCI and Software Engineering through Risk Analysis

TL;DR

This research addresses the "coordination paradox" in distributed design teams. As software systems grow more complex, teams lose the "big picture." The authors propose a methodology that integrates Risk Management directly into the Claims Analysis (Design Rationale) process. By treating design trade-offs as quantifiable risks, the LINK-UP system helps teams maintain a collective memory and prioritize the most critical tasks during iterative redesign cycles.

Problem & Motivation

Software development is increasingly distributed across space and time, yet existing tools like SOPPTS (task-oriented) or TeamSCOPE (awareness-oriented) often add excessive overhead or provide too much low-level data without strategic context.

The authors argue that human-computer interaction (HCI) and project management share a common DNA: they both combine technical constraints with human psychological concerns. The core problem is that design knowledge is often lost, and teams take an ad hoc approach to what should be a scientific, engineering-driven process.

Methodology: Risk as the Engine of Design

The heart of this work is the extension of Carroll’s Claims Analysis. Traditionally, a "claim" documents the upsides and downsides of a design feature.

1. From Product Claims to Process Claims

The authors introduce a novel distinction:

• Product Claims: Related to the notification system's interface and utility (e.g., "Non-intrusive alerts improve focus but might be missed").
• Process Claims: Related to the team's methodology (e.g., "Weekly syncs improve coordination but reduce development time").

2. The Risk Model

By treating the downsides of these claims as project risks, the authors apply a Software Engineering risk model to HCI. Every potential problem is assigned a weight based on:

• Impact: How devastating the downside is (validated by theory or previous experiments).
• Probability: The confidence level that this downside will occur, adjusted by mitigation efforts.

Figure 1: Comparison of Product vs. Process Claims and their associated downsides (Risks).

The LINK-UP System: Tool Support

The methodology is implemented in LINK-UP, a web-based suite designed to guide usability engineering. Key features include:

• Prioritized Redesign: After an evaluation phase, the system identifies "broken" claims and ranks them by risk exposure, telling the team exactly what to fix first.
• Activity Timeline: Exploits episodic memory by visualizing design changes and team decisions over time, preventing "knowledge rot."
• Knowledge Reuse: Teams can pull claims from a repository of previous projects, ensuring that "mistakes are made only once."

Figure 2: The architecture for supporting collaborative design through externalized team memory.

Experiments & Results

The paper emphasizes that effective team coordination requires shared knowledge—knowing that everyone else knows the same thing.

• Awareness vs. Overload: By using notification systems and time-based filtering, the authors show how teams can monitor teammate activity without being distracted from primary tasks.
• Redesign Guidance: LINK-UP transitions from a simple documentation tool to a "guide" that structures subsequent design iterations based on empirical evaluation results.

While the paper is a conference contribution (ACM SE05) focused on the framework's architecture, its primary "result" is the successful integration of SE risk disciplines into HCI design cycles—a precursor to modern "DesignOps."

Critical Analysis & Conclusion

Takeaway

This work provides a roadmap for Science of Design. It suggests that project management shouldn't be a tedious administrative layer; it should be the logical outcome of documenting design rationale.

Limitations

• Overhead Concerns: Although the authors claim minimal overhead, maintaining a robust claims library requires significant discipline from the design team.
• Automation: The 2005 context lacks the modern AI capabilities we have today to automatically extract claims from Slack or GitHub, which would make this system much more scalable.

Future Outlook

As we move toward AI-assisted design, the concept of a Knowledge Repository for claims could be the perfect training ground for LLMs to act as "Risk Consultants" for human design teams.