Pragmatic vs. Perfectionist Approaches in Knowledge Modeling

Overview

This article explores the tension between perfectionist and pragmatic approaches in knowledge modeling, drawing from observations about the evolution of knowledge representation frameworks from the 1990s to present day.

Historical Context

Following themes in Holger Knublauch 's research (see below and see also Holger Knublauch) reveals well-established concepts that have remained largely unchanged. The Generic Frame Protocol Specification (GFP) specification from 1997 provides illuminating insights into terminology such as "Slot," which remains common in modern frameworks like LinkML. Structurally, the field has seen limited evolution, with few approaches successfully bridging fundamental conceptual divides.

The Person-Address Modeling Problem

A classic example illustrating the complexity of knowledge modeling involves the seemingly simple relationship between a person and their address:

Initial Naive Assumption

The initial modeling approach assumes a functional dependency: knowing a person's ID automatically provides their address.

Reality Complications

Real-world scenarios reveal multiple layers of complexity:

Identity vs. Documentation: The need to distinguish between a person and their identity documents
Multiple Identities: Exotic cases including:
- Intelligence operatives with multiple official identities
- Witness protection program participants
- Criminals using falsified documents

Address Type Proliferation

Addresses prove to be constructs serving different purposes:

Delivery addresses - for reliable postal service
Billing addresses - for financial transactions
Residential addresses - actual living locations
Vacation addresses - temporary locations
Digital nomad addresses - challenging the concept entirely

Two Philosophical Approaches

Perfectionist Approach

Advocates for:

Precise modeling of all edge cases
Comprehensive agreement and clarification processes
Quality improvement through detailed specification
Aspiration toward theoretical perfection

Pragmatic Approach

Recognizes:

"Good enough" solutions as inevitable
Statistical and economic constraints
Organizational inertia and established practices
Operational validation over theoretical completeness

Frustration-Avoidance Methodology

Core Principle

Rather than pursuing optimization, the frustration-avoidance approach prioritizes:

Frequency-based relevance assessment
Statistical relationships over rigid constraints
Use case-specific modeling decisions

Cardinality Handling

Traditional rigid cardinalities are replaced with statistical relationships to use cases:

Use Case A: Always true
Use Case B: Frequently true
Use Case C: Rarely true
Use Case D: Never true

Implementation Strategy

Use case-specific APIs and queries through targeted filtering
Local subgraph optimization for specific scenarios
Skepticism about global optimization approaches

External Links

Case Study: Using Prot´eg´e to Convert the Travel Ontology to UML and OWL

⚠️ LLM-generated content notice: Parts of this page may have been created or edited with the assistance of a large language model (LLM). The prompts that have been used might be on the page itself, the discussion page or in straight forward cases the prompt was just "Write a mediawiki page on X" with X being the page name. While the content has been reviewed it might still not be accurate or error-free.

Pragmatic vs. Perfectionist Approaches in Knowledge Modeling

Contents