What I’m interested in

Everything on this page flows from one professional imperative: contestability. The job, for most of a career, has been to test someone else’s case — spot the gap, challenge the flawed step, and understand why the answer is what it is. Why does a navy need nine frigates, and not three, and not fourteen? Nine is not a taste. It falls out of structure: the arithmetic of rotation that turns nine hulls into roughly three at sea, concurrency, maintenance, crewing, the tasks a government has actually signed up to. A challenge that engages that structure is worth hearing. A challenge that just doubts the number is noise.

The part that shapes everything else: the contest comes in from outside. I had not done naval warfare before contesting naval capability, and that is not a defect in the arrangement — it is the arrangement. A sound case must be logically consistent to an intelligent outsider. “Trust me” is not a defence, and neither is “come back after twenty years in the domain”; both are findings. If the structure that generates the answer cannot be walked by someone outside the priesthood, that gap is the report.

Which forces a method. Coming in cold, I have to be able to understand the world, break it into parts, understand the flows and constraints, and test whether what is being said matches what actually happens. The four interests below are that sentence taken one clause at a time. None of them arrived as a hobby; each got built because the contest required it.

Before a claim can be contested you have to know precisely what it says — and most contested claims turn out, under pressure, to be vague about their own parts. Ontology is the discipline of not letting that slide: deciding what the things are, what counts as an instance of each, how they relate, and then committing to the vocabulary and living inside it. The commitment is the hard part. Anyone can draw a taxonomy; the discipline is refusing to let a term quietly mean two things when one of them would be convenient.

It is also where the hidden assumptions live. Pin down what “frigate” means for the purposes of the case — a hull, a crewed and worked-up ship, a deployable unit of presence — and three apparently contradictory force sizes become answers to three different questions. Most disagreements I have been paid to resolve were ontological before they were analytical: the parties were not disputing the number, they were using the same word for different things.

The index keeps circling this interest for a reason. Controlled vocabularies — committing to a set of terms and then having to live inside them — are the same discipline pointed at software, where the machine is the intelligent outsider the definitions have to survive.

A decomposed system is still a photograph. Systems thinking is what happens when you let it run: stocks, flows, feedback and delay — the machinery by which a change over here becomes a consequence over there, later, with the sign flipped. It earns its place in the method through the counter-intuitive flow: the intervention that nominally improves something and actually makes it worse, because its second- and third-order effects were not in the picture the decision was made on.

Every experienced reviewer keeps a collection. Add capacity to the congested system and induce the demand that congests it further. Tighten a reporting requirement to get honesty and watch it optimise the reports instead of the behaviour. The shape is always the same: the case was argued on the first-order arrow, and the loop was left off the diagram. So the contest is largely the practice of asking where the loop is — what this change does to the thing that feeds it.

The lens also names the fix: leverage. Most interventions push where the system is stiff; occasionally there is a point where a small change propagates. Finding it requires the map — which is why this interest and the previous one are really one interest. You cannot trace flows between parts you have not named.

No claim lands in a vacuum. Between the proposal and reality sits everything already built: the systems it must talk to, the standards it must satisfy, the commitments — technical, contractual, political — made years ago by people solving different problems. Enterprise architecture is the discipline of fit at that scale, and it produces a verdict the other lenses cannot: technically fine, architecturally wrong. The thing works; it just cannot live where it will have to live.

It is also where systems carry their history most visibly. An enterprise is an evolved organism: environment and constraint drive form, form constrains behaviour, and the whole thing hauls its ancestral commitments along as baggage it cannot cheaply put down. Reading a proposal architecturally means reading it as a time series — not “is this good”, but what does it inherit, what does it collide with, and what will it commit us to that someone will be contesting in ten years.