For when rants trigger RCU stall detection

and my patience buffer overflows into /dev/null. The rants are technically supposed to be brief. And let’s be real, some system bugs deserve a full kernel panic’s worth of fury.

A world map in the muted, slightly aged style of an old commercial atlas, deep greys and warm parchment tones, no bright colours. The landmasses are plain and understated, deliberately unremarkable. The eye is drawn instead to the chokepoints: a handful of narrow sea passages marked with subtle bright constrictions, like a ribbon pinched between two fingers, at the Strait of Hormuz, the Strait of Malacca, Bab-el-Mandeb and Suez, the Panama Canal, the Danish and Turkish straits. Thin flowing lines of shipping traffic stream across the oceans and visibly bunch and narrow as they squeeze through each pinch point, thick and tangled at the straits, thin elsewhere. At several of these pinch points sits a small, discreet military marker, a simple anchor or star symbol with a faint defensive ring around it, showing a base or fleet holding the passage: one at Hormuz, one in the middle of the Indian Ocean at Diego Garcia, one at the mouth of Bab-el-Mandeb near the Horn of Africa where several such markers cluster together uneasily, one at Panama. Over the landmass of China, a different marker: not a passage but a cluster of small refinery and smelter symbols glowing faintly, with dispersed thin ore-lines from Africa, South America and Australia all converging into that single cluster, showing the scattered sources funnelling into one processing hub. Across the Arctic top of the map, a faint dotted route curves over the pole, bypassing the pinch points below, marked discreetly as the northern passage, with a single marker of its own. In one corner, almost incidental, a few small human figures rendered very small and plain, standing apart from the grand map, suggesting the people downstream of the passages who never appear on it.

The map is mostly bottlenecks

The usual way of reading the contest over resources is to look at where they are. Whose ground holds the oil, the gas, the rare earths, the cobalt. On that picture power follows the deposit, and the country sitting on the richest seam holds the strongest hand. It is the picture behind a great deal of the current talk about resource grabs, and it is wrong in a way that the talk mostly misses. ...

A lone figure seated on an ornate throne-like chair at the centre of a grand hall, gripping the armrests with the confident posture of someone who believes the room revolves around them. The chair stands on a circular dais, but the dais is wheeled and slightly tilted, beginning to roll away unnoticed. All around, smaller groups of figures have turned their chairs to face one another in tight conversational clusters, forming their own little circles, none of them looking at the central figure. A few have already drifted toward the open doors, where soft daylight spills in. The central figure holds a length of rope that once tethered the surrounding chairs, but the rope is slack and the knotted ends have quietly come loose on the floor.

The centre that depends on the room

The usual way of reading power treats it as a possession. Someone holds the leverage or they do not; the troops, the money, the umbrella, the seat at the head of the table. On that picture the strong actor is the one who can hurt the others more than they can hurt back, and the others arrange themselves accordingly because they have no better option. It is a tidy model, and it is wrong in a particular way that takes a while to show. ...

A shape of European rearmament

Everyone knows by now that Europe is rearming. That part is easy to write and carries almost nothing, because “investing in defence” survives any amount of vagueness. The interesting reading sits a little lower down, in the documents that have to name things: how much money, raised in what way, spent on what, through which legal route, under whose command. Those are harder to write and harder to take back, because somebody has to act on them. Read side by side, they sketch a shape, and the shape says more than the announcements do. ...

Reading the procurement, not the press release

A policy speech says what an institution wants to be seen wanting. A procurement document records what it has agreed to pay for. The two are not the same, and the gap between them is usually the more interesting reading. A press release costs nothing to write and binds no one. A call for tender names a budget, a deadline, a set of deliverables and a contract that someone will sign and then have to deliver against. Intentions are cheap. Commitments leave a paper trail, because somebody has to do the work and somebody has to account for the money. ...

A single weary civil servant at a cluttered government desk, buried under towering stacks of paper documents and redaction folders, frantically rubber-stamping and blacking out pages. In the background, an empty pristine filing system labelled 'intake' sits untouched and gathering dust.

Een transparantiewet die om vindbaarheid vraagt

Er is een gewoonte in detection engineering management die een organisatie tijd kost om af te leren. Een team kijkt naar de waarschuwingen, de dashboards, de dekkingsrapporten en behandelt ze als de plek waar detectie gebeurt. Dat is niet zo. Het is de plek waar detectie zichtbaar wordt. Het werk dat bepaalt of er überhaupt iets te zien valt, gebeurde veel eerder en veel stiller toen iemand koos wat er gelogd werd. Op een event dat nooit is vastgelegd valt geen alert te schrijven. De detectie aan de uitgang reikt niet verder dan de telemetrie die iemand aan de ingang heeft ingericht en niemand voelt de prijs van een ontbrekende log entry op het moment dat de log niet wordt geschreven. Die prijs komt later, tijdens een incident, wanneer blijkt dat het te reconstrueren spoor nooit heeft bestaan. ...

The stability of dysfunction

The stability of dysfunction Many discussions of large systems quietly assume that a stable system is a healthy one, and an unstable system is a sick one. Complex systems tend to violate that intuition. They can remain operational for a very long time without becoming any healthier, and they possess several mechanisms that let them do so. Normalisation is one of those mechanisms. It is not the whole story. The whole story, if there is one, is that persistence and health are different properties, and large systems are often much better at achieving the first than the second. ...

What institutions do to successful ideas

Many critiques of Agile assume something went wrong. The story is usually told as a fall from grace: a practical response to software uncertainty that then disappeared beneath ceremonies, certifications, frameworks, and consultants. There is another way to read the same history. Perhaps Agile did not fail. Perhaps it succeeded, and what happened next is simply what institutions do to successful ideas. On that reading the interesting question is not why Agile became institutionalised. It is why successful ideas so reliably do. ...

Quadrant mapping uncertainty against consequence, with adaptation favoured at high uncertainty and anticipation at high consequence

Agile, where it fits and where it doesn't

Most engineering disciplines accept that method follows from context. Nobody expects a bridge engineer, a documentary director, and a trauma surgeon to share a planning model, and nobody finds the difference remarkable. Software is one of the few fields where people go looking for a single methodology and then try to apply it to everything in sight. Agile is the most successful instance of that search, which makes it an awkward thing to argue about. The useful question is not whether it is good or bad. It is what kind of work it was meant for, and that turns out to depend on two things that have nothing to do with how much anyone likes the method. ...

How a rebellion became a bureaucracy

Few movements in software have been as successful as Agile. What began as a reaction against heavyweight process, exhaustive documentation, and centralised planning became the dominant way organisations talk about building software. And somewhere in that success it acquired certifications, prescribed ceremonies, maturity models, governance structures, a consulting industry, and dedicated management hierarchies, which is to say it acquired most of the things it was a reaction against. The question worth asking is narrow. Not whether Agile succeeded, and not the broad point that institutions reshape ideas, but the specific mechanism. By what steps does a critique of bureaucracy turn into a bureaucracy, when nobody involved wanted that outcome? ...

A figure at a small desk in a high-ceilinged reading room, reading with the quiet satisfaction of someone who has found exactly the book they wanted; through the tall windows, crowds of other figures drift past, half-dissolving into a cheerful drizzle, their attention already elsewhere. The reader is entirely unbothered.

Audience design

A friend asked whether the proof-of-concept I had sketched on a docs page was worth actually building. I said no, for the usual reasons, and also because the page in question had been written in the specific style of a document that does not want to be read. A four-layer architecture diagram without the diagram. Ingestion, storage, correlation, presentation. Bullet lists of API names with the dispiriting authority of a railway timetable. It reads like a tender response written by a tender response. ...