Playbook:

The Operating Context System Playbook

A blueprint for building the context infrastructure that makes organizations legible to themselves and to the AI systems working alongside them.

Chapter 11

The Boundaries of Structure

Minimum viable context and the danger of ontology bloat

Part X: The Boundaries of Structure

Over-modeling is as dangerous as under-modeling.

One of the most common failure modes for Operating Context Systems is ontology bloat: the gradual accumulation of artifact types, relationship types, and properties until the model becomes too complex to maintain and too rigid to adapt.

There's a tempting logic that more structure is always better. This is wrong. Every artifact type added to the model is a maintenance obligation. Every relationship type creates complexity in how the graph is queried and maintained. Every property added to an anchor artifact is a field that someone (or some agent) needs to keep current.

Structure should earn its place. The test is whether the artifact type, relationship, or property is necessary for a decision or insight that matters. If it isn't driving anything, it's just overhead.

Question to AskIf YesIf No
Is this type of thing referenced repeatedly across decisions and plans?Make it a first-class artifactLeave it as a tag or attribute
Is the information too important to live in unstructured documents?Structure itLeave it as a document
Is there a clear owner for keeping this accurate over time?Add it to the modelDon't model it yet
Will this go stale quickly without active maintenance?Build a decay mechanism before adding itProceed with standard maintenance

The right frame is "minimum viable context": not the smallest possible operating model, but the smallest one sufficient to support the planning and decision-making the organization actually needs to do.

Next

Continue reading

Interoperability: The Open Context Layer