The first time I wrote that sentence, it pointed at workflow. The site had stopped being only a place where pages lived and had become a system that could keep public output from falling apart. Timestamps, series links, corrections, uploads, and a command instead of manual repetition were not glamorous changes, but they mattered because they reduced the friction between having work and actually placing it somewhere visible.
That was the first shift.
This is the second one.
Now the sentence points less at workflow and more at the public surface itself. Hedegreen Research no longer behaves truthfully as one flat website with a growing archive inside it. It has started to split into rooms, and that split is not there because bigger looks impressive or because every line of work deserves its own branded corner. It is there because different kinds of public work eventually need different doors if the structure is going to stay honest.
One Archive Was Enough
At the beginning, one archive really was enough. Articles could live under articles, builds could live under articles, corrections could live there too, and open notes could live there as well. That was not a mistake. It was the right first layer, because the problem then was not classification but friction. The work needed somewhere public to land without each piece becoming a custom publishing event, and the archive solved that problem.
It still does.
The archive remains the spine.
But a spine is not the whole body.
At some point the work stopped being mostly prose. It started producing tools, instruments, rooms, reports, sound, and small public utilities that were not best understood as essays about themselves. Once that happened, one flat archive stopped being a complete map.
Different Work Needs Different Doors
The issue was not that there were suddenly more pages. The issue was that the public actions had become more varied.
An article is read.
A tool is used.
A correction is checked.
A room is entered.
A report is received.
A song is heard.
Those are not the same public action, and if they all arrive through the same kind of door, the structure may look simpler from a distance while becoming easier to misunderstand up close. That is why the split matters. The archive explains. TID holds instruments and data artifacts. Signal holds sound and room logic. The public repos can hold mechanism when mechanism itself matters.
Each surface has a different job.
The work did not become clearer by being flattened.
It became clearer by being placed.
The Shell Changed The Proof
V2 made that easier to see because the claim no longer rests only on naming. Before that, an article could still feel like an isolated page with some decorative chrome above it. You entered, you read, and you left, and even if the project behind the page was already changing shape, the page could still present itself like a sealed island.
That is less true now.
The shell can stay alive while the room changes.
Time can remain present.
Signal can remain present.
Local can exist as a reader-side layer without pretending to be a generic settings box.
None of that proves the structure through decoration. It proves something narrower and more useful: the public frame itself has started carrying room logic. An article no longer has to act like a detached document hanging under dead page chrome. It can sit inside a place that continues around it.
That matters because a claim like “this stopped being just one website” should not rest only on directory names or future plans. It should be visible in use. The shell is now part of that proof.
TID, Signal
TID is a good example of why naming alone is not enough. From outside, /tid/ can look like just another path, or a strange label, or a miscellaneous bucket with a dramatic name. Inside the system, that is not what it is. TID is the public instrument layer. It is where things live when they need to remain reachable as artifacts rather than only described in prose: opened, checked, returned to, and used again.
That is why TID is not a category bucket.
It is a room.
The door stays stable even when the objects behind it vary.
Signal forced the same distinction from another side. The music line could have been handled only through articles, but that would have been wrong. A song is not only an argument about a song, and a room is not only the essay explaining why the room exists. Signal holds sound, characters, internal mythology, and the moving edge between music and system. It has articles around it, but it is not reducible to those articles. Once that became true, Signal needed a different public door.
The Guardrail
There is a real danger here. Every project can invent a label, every label can demand a page, and every page can pretend to be a room. That is how systems become unreadable.
So the rule cannot be:
make more rooms.
The rule has to be stricter than that. Make a room only when the work type requires a different public action. If the visitor reads it, it may belong in the archive. If the visitor uses it, it may belong in TID. If the visitor enters it as sound and room logic, it may belong in Signal. If the visitor needs visible mechanism, the code may belong on GitHub.
Multiplicity is only discipline if it reduces confusion.
Otherwise it is vanity with directories.
The Better Name
That is why the old title still works, but differently now. This stopped being just one website not because the homepage became bigger, not because there are more links, and not because the domain changed, but because the public system now has to carry different behaviors honestly: reading, using, checking, listening, receiving, returning, and following a trace from article to instrument to report to source.
One flat surface cannot carry all of that forever without starting to misdescribe the object.
The answer is not complexity for its own sake.
The answer is legibility.
Each door should tell the visitor what kind of object they are approaching. Each room should justify its name by what it holds. Each split should make the system easier to understand rather than harder.
That is the test that matters to me. If the rooms make the work clearer, the split was structural. If the rooms only make the project look larger, the split was vanity.
Hedegreen Research has to stay on the first side of that line.
The archive remains the spine.
The shell keeps time and system presence visible.
TID holds instruments.
Signal holds sound.
The public surface is becoming a room-family. That is not a slogan so much as a maintenance problem and a naming correction at the same time. It is a better description of what this thing has become, but only if the structure keeps earning the name.
— Dennis Hedegreen
