-
The Operating System of Governance: Policies and Standards Done Right
Governance isn’t about control but about clarity,n responsibility and embedded support. When roles are clear, metadata drives applicability, and platforms make compliance effortless, governance becomes infrastructure and not friction. This post offers a modern, domain-driven take on making that real.
-
Purpose-Driven Interfaces in the Data Mesh Era
A single data product rarely serves all needs through a single interface. Exploration, integration, and training demand different guarantees—of freshness, stability, and performance. Purpose-driven interfaces make these differences explicit. They’re not just technical entry points, but contracts shaped by intent, reinforced by governance, and defined through service level expectations.
-
If You Want Discoverability, Don’t Ditch JSON Yet
Many see REST and JSON as inseparable—but they’re not. REST is a design style; JSON just happens to be where discoverability lives today. If you want semantics, navigation, and affordances—especially for AI agents—you still need JSON. Not because it’s ideal, but because no other format supports it yet.
-
Keep the Knowledge Graph Light: Keep the Compass, Drop the Cargo
If you’re building a knowledge graph (KG), you might be tempted to pour all your data into it, including time series. After all, they’re important. They power analytics, forecasts, anomaly detection, and digital twins. But storing time series in the KG is almost always the wrong move. You don’t need to, and probably shouldn’t, put…
-
Beyond the Monolith: Why Federated Knowledge Graphs Matter
Organizations increasingly recognize that centralized semantic approaches do not scale, as departments have distinct languages. The Federated Knowledge Graph (FKG) model promotes local ownership of semantics, allowing diverse graphs to coexist. This model encourages interoperability through mapping layers and supports AI by providing context, enabling reasoning across disparate systems without enforcing uniformity.
-
More Connections, Not More Data: What a Knowledge Graph Really Is!
A knowledge graph fundamentally transforms data management by emphasizing relationships over isolated facts. It connects entities through meaningful edges, enabling systems to understand context and reason efficiently. Unlike traditional data systems that prioritize volume, knowledge graphs focus on creating structured insights, reflecting the importance of connections akin to neural networks in the brain.
-
The Custodian, the Lift Operator, and the Visitor: Understanding Data Systems by Role, Not Tool
In modern data architecture, effective system management relies on defined responsibilities among components: the database (custodian) maintains data integrity, the pipeline (lift operator) prepares data for use, and the query engine (visitor) explores it. Clarity in these roles fosters trust and effectiveness, crucial for resilient data systems.
-
Not Everything Needs to Think: Where RPC, GraphQL, and Other API Styles Still Matter
The discussion examines the relevance of RPC, GraphQL, and events alongside newer API paradigms like Intent and Planning APIs. RPC excels in speed and precision for internal calls, GraphQL provides flexible data retrieval, and events facilitate system observability. A layered architecture combining these approaches enhances system efficiency and responsiveness.
-
Beyond REST: Intent and Planning APIs for an Agentic World
The article discusses the evolution of REST with the introduction of Intent and Planning APIs, enhancing interaction models for both human and AI clients. These APIs shift focus from traditional imperative workflows to declarative goal-oriented interactions, allowing systems to plan and propose outcomes, fostering transparency and adaptability in modern applications.
-
REST Reborn: From Integration Layer to Decision Interface
People keep asking if REST is dead. No, but for years, it was underused, misunderstood, and often reduced to CRUD-over-HTTP. What is dead (or should be) is the naive REST of thin wrappers, flat links, and humans in the loop.REST with no hypermedia. No semantics. No introspection: Zombie REST. But here’s the twist: the features…
Ambiguity vs Information
A journey in a world made of data, information, knowledge and wisdom.
Ambiguity vs Information