Rotterdam in the 1980s and 1990s did not treat colour as a polite accent.
Colour was on buildings, posters, museum walls, school façades, tram stops, festival graphics, children’s books, public art, and the civic furniture of the city. Red could be red. Blue could be blue. Yellow did not have to ask permission from beige. Shapes were allowed to be simple, blunt, cheerful, and a little unruly. Blocks of colour did not apologise for taking up space.
I did not have language for that as a child. I did not walk around Rotterdam thinking about Dutch modernism, CoBrA, post-war reconstruction, or the movement of visual ideas between the Netherlands, Denmark, Sweden, and Finland. I just knew the city had a certain grammar: hard edges, bright fields, public confidence, and a refusal to make everything tasteful by turning it grey.
Karel Appel was part of that weather. Not as a tidy style reference, and certainly not as a design system. He was louder than that: more bodily, more improvised, more hand than grid. His work had the same trust in direct colour, but it arrived through force and play rather than proportion. Appel feels, to me, like the organic cousin of the more architectural Scandinavian register the site has ended up using: the same permission structure, but with different muscles.
That connection helped explain why the Stockholm Design Lab direction felt less like a costume than the earlier attempts. We had visited a few stations on the way here: literary paper, cabinet language, antique texture, clever little frames. They all had charm for a minute. None of them felt native. They made the site look like it had found a pleasing outfit rather than a real house style.
The Swedish and Finnish direction changed the question. Instead of asking whether the site looked literary, it asked whether the work could stand inside a clear visual world: saturated colour fields, direct labels, heavy sans type, rows, ledgers, hard edges, drawings that behave like systems, and enough white space for the whole thing to breathe. That vocabulary can hold software, cloth, walking, memory, DevRel, taste, and agents without pretending they are the same subject.
The missing piece was not more grid. It was permission for the illustrations to have a body.
That is where Appel found a home. The headers can remain structural: Scandinavian colour fields, measured routes, hard proportion, the page standing up straight. The images inside the articles can do a different kind of work. They can keep the same palette and page ground, then let the forms become more organic: rough black contours, childlike painted shapes, textile density, awkward overlaps, and the small expressive violence that makes an idea feel handled rather than merely diagrammed.
That freedom has to earn its usage. A wild mark is not there to decorate the page. It has to make the concept more inspectable. In the clothing essays, the mark can carry cloth, weight, pocket, weave, and the way a jacket changes the body. In the technical posts, it can turn a protocol, a cache, a test, a permission boundary, or a rollback path into an object with pressure and consequence. The Scandinavian grid keeps the page from shouting. The CoBrA impulse keeps the page from becoming bloodless.
It also has roots I recognise. The Stockholm version is more disciplined than the Rotterdam memory. Less expressionist, less scraped and flung. More wayfinding, identity, institution, architecture. But the family resemblance matters. The colours are not decorative. The shapes are not timid. A page can be a field. A title can be a structure. A list can be a civic object, not a stack of little content cards.
Once that became clear, the writing had to move with it.
The prose could not keep acting like a small inherited journal. It needed the same directness as the design: concrete incidents, named objects, firm claims, no theatrical dust. A technical post should begin with the tool, failure, check, or boundary that changed the work. A personal essay should begin with the place, garment, weather, route, room, or memory that actually caused the thought. The lesson comes later, if it has been earned.
The site now has a better test for every page: does the design help the material become more inspectable, or is it asking the material to perform a mood?
That test catches the old mistakes quickly. A paper texture can make weak prose look considered for about thirty seconds. An antique caption can give a paragraph borrowed gravity. A faux archive can make ordinary indecision feel curated. These devices are attractive because they offer atmosphere without requiring judgement. They also age badly because they are props.
The current direction is less forgiving, which is part of the appeal. A giant word on a colour field has nowhere to hide. A row of projects either explains the work or it does not. A technical post either has the receipt trail or it becomes a handsome list of claims. An essay either notices the actual thing in front of it or it becomes taste vocabulary looking for a room.
Dutch colour memory gives the site warmth. Scandinavian structure gives it manners. The two together feel closer to the right public voice: direct, saturated, useful, unsentimental, and still capable of pleasure. Appel no longer has to stay in the background. He can live in the article images, where expression has a job: adding bodily texture, unruly intelligence, and lived material force to a strong Scandinavian grid. Stockholm provides the restraint that keeps the page from becoming a shout.
Category-crossing is where the decision earns its keep. A site about coding agents and oxford cloth can easily look incoherent if the design keeps changing costume. It can also become bland if the design tries to average the subjects into neutrality. The house style has to do something harder: give each subject enough space to be itself while making the whole place recognisable.
For the technical writing, that means ledgers, systems, exact titles, and claims with evidence underneath them. For the essays, it means memory, cloth, weather, streets, taste, and private rooms handled with enough discretion that the reader is not asked to become a confidant. Both modes can share a typeface, a palette, a grid, and a manner of attention. They should not share a false mood.
Names still matter when they carry evidence. Rotterdam matters because it explains why these colours feel familiar rather than newly adopted. Karel Appel matters because he names one organic branch of the same permission. Stockholm Design Lab matters because it names the disciplined branch: identity, proportion, wayfinding, type as architecture. The useful part is not prestige by association. The useful part is recognition.
By that measure, the house style is not a costume. It is a way of refusing costumes.
No literary dust. No cabinet. No terminal. No little tricks to imply authority. The site gets its authority from the work it can show, the claims it can support, the objects it can name, and the confidence of its visual decisions.
Colour can be colour. Type can be type. A page can stand up straight.