Uncovering the Fundamental Purpose

• This presentation uses the Design & Engineering Methodology for Organisations (DEMO) to uncover the fundamental purpose of the Semantic Web.
• At its core, the Semantic Web facilitates the creation of new, verifiable facts on a global, distributed scale.
• DEMO distinguishes between the what (ontological essence) and the how (performative implementation) of an enterprise.

A Note on 'Ontology' in DEMO

• In the Semantic Web: An ontology is a formal, explicit specification of a shared conceptualization—a data model for classes, properties, and relations (OWL, RDF) intended for machine use.
• In DEMO: An ontology refers to the Enterprise Ontology, a precise scientific notion of an enterprise's essence. It focuses on fundamental human acts of communication and commitment that bring new facts into existence—the "what" rather than the "how".

What is the Fundamental Nature of an Organization's 'Being'?

• According to DEMO, an organization's being lies in the people who enter into and comply with commitments.
• This essence consists of coordinated acts between actors that create new facts—the building blocks of any enterprise.
• This ontological core is stable, independent of changing technologies, processes, or structures.

The DEMO Lens: Actors, Facts, and Transactions

• DEMO asserts that organizations create ontologically distinct facts through structured communicative transactions.
• A minimal transaction pattern exists:
  • An initiator requests a result from an executor
  • The result is declared
  • The result is accepted
• The essence is the commitment of actors to bring a new fact into being.

The Essence: An Ontological Verification Transaction

• We can model the Semantic Web's essence as a single fundamental DEMO transaction: an ontological process of knowledge creation and verification.

Transaction Components

• Transaction Kind (TVerification): Verifying a claim.
• Actor Role (Initiator): The A_Claimant, who asserts a fact (e.g., "The capital of France is Paris").
• Actor Role (Executor): The A_Verifier, who declares the truth or falsity of the claim. In the Semantic Web context, this verifier is distributed—a network of sources, agents, or consensus mechanisms.

The Verification Transaction Phases

Order Phase

• The claimant makes a communicative request to the verifier.
• The claim is the product (PClaim).

Execution Phase

• The verifier queries distributed sources and applies rules to assess the product.
• The knowledge base is dynamic and formed through linked data.

Result Phase

• The verifier promises and declares the result.
• If confirmed, the product becomes a P_VerifiedFact.
• If rejected, exception handling follows.

The Other Half of the Value Chain: Consumption

• DEMO emphasizes that transactions are incomplete until the initiator accepts the result—the consumption act.
• The newly created P_VerifiedFact becomes an input for the claimant's subsequent actions.

Example

• A software agent may request verification of a product spec.
• After receiving a P_VerifiedFact, the agent uses it to drive an ordering or compatibility check.
• Without consumption, verification is merely academic.

Uncovering the "Problem"

• Through the DEMO analysis, the Semantic Web's purpose becomes clear: it provides the technological infrastructure for the TVerification transaction, particularly across distributed, decentralized knowledge sources.
• It solves the absence of a scalable, machine-interpretable, decentralized mechanism for knowledge verification.
• Traditionally, verification is a slow human activity; the Semantic Web automates and scales this ontological act.

A DEMO Critique of Semantic Web Practice

• From a DEMO perspective, the Semantic Web community may have overemphasized the how (RDF, OWL, SPARQL) at the expense of the what.
• While these implementation details are necessary for interoperability, their triple-based structure can sometimes limit the expressiveness needed to capture complex business rules and constraints.
• A modeling-first approach, which focuses on a richer and more expressive conceptualization before committing to a specific technical representation, offers a potential way to mitigate this.

Closing Thoughts: LLMs, Provenance, and a Hybrid Future

• Large Language Models (LLMs) such as ChatGPT and Gemini present a challenge: they can synthesize unstructured information efficiently, yet their outputs often lack provenance and accountability.
• A promising solution is a hybrid approach: using LLMs as instruments for both claimant and verifier while keeping final responsibility for declarations with accountable humans or authorized agents.