Every dimension the house is generated from is below, with its class and its citation. This is the project’s honesty layer, and it is given a room of its own because a smooth 3D render implies a precision the sources do not have.
The dimensions
| Dimension | Value | Class | What it means |
|---|---|---|---|
| 2.60 mIf 20% out: 2.60 m | Interpolated | Length of one bay along the ridge. The length of the house is this figure times the number of bays, and that number is always odd — the house is named by it. It also has to be large enough that even the shortest house stays longer than it is wide: a building lying north–south is not a rumoh Aceh, whatever shape its roof is. | |
| 2.20 mMoves no measurement | Interpolated | Depth of the seuramoë keuë, the front veranda where guests and the men are. | |
| 2.90 mMoves no measurement | Interpolated | Depth of the tungai, the raised middle room. This is where people sleep and where they are born; it is the most closed part of the house. | |
| 2.00 mMoves no measurement | Interpolated | Depth of the seuramoë likôt, the back veranda where the women work and cook. | |
| 0.35 mIf 20% out: 0.07 m | Interpolated | Height of the tungai floor above the two verandas. The Palembang rumah limas raises its floor to state a guest’s standing; here what is raised is the most closed room, and what is stated is something else entirely: not who is higher, but how far an outsider goes. | |
| 2.50 mIf 20% out: 2.00 m | Interpolated | Height of the veranda floor above the ground. High, and the space beneath is used for work, storage and weaving. | |
| 0.06 mIf 20% out: 0.01 m | Interpolated | Thickness of the board floor. | |
| 0.20 mMoves no measurement | Interpolated | Section of a tameh, the post standing free on its footing. | |
| 1.60 mMoves no measurement | Interpolated | Spacing of the lines of posts across the width. | |
| 0.26 mMoves no measurement | Interpolated | Depth of a toi, the beam threaded through a mortise cut clean through the post. This frame is lashed and pegged with no nails, which is what lets it move without breaking. | |
| 0.14 mMoves no measurement | Interpolated | Width of a toi. | |
| 2.10 mIf 20% out: 0.42 m | Interpolated | Height of the board wall from the floor to the eave. | |
| 0.04 mMoves no measurement | Interpolated | Thickness of a wall board. | |
| 0.29 mIf 20% out: 2.00 m | Interpolated | Rise of one tread. This figure sets how many treads there are, and that number has to be odd — so nudging the rise can flip odd to even without changing anything else. It is what this building’s counterexample pushes, and the smallest counterexample in the project. | |
| 0.27 mMoves no measurement | Interpolated | Depth of a tread. | |
| 0.95 mMoves no measurement | Interpolated | Width of the ladder. | |
| 3.10 mIf 20% out: 0.62 m | Interpolated | Rise of the ridge above the eave. | |
| 1.10 mMoves no measurement | Interpolated | Depth of the overhang. | |
| 0.08 mMoves no measurement | Interpolated | Section of a rafter. | |
| 0.14 mMoves no measurement | Interpolated | Section of the eave plate. | |
| 0.22 mMoves no measurement | Interpolated | Exposed depth of one course of sago thatch. | |
| 0.04 mMoves no measurement | Interpolated | How far a course stands proud of the one below. | |
| 0.50Moves no measurement | Interpolated | The share of a course the course above laps. | |
| 0.04 mMoves no measurement | Interpolated | Clearance between the frame and the first course. | |
| 0.30Moves no measurement | Interpolated | Depth of the tested joint engagement, as a share of the smaller member. | |
| 1Moves no measurement | Canon | The ridge lies east–west, because prayer is toward the west: the room is already turned that way before anybody stands in it. The other nineteen buildings in this project are oriented by something local — a compass rule of their own, a granary, a river, a road, a stone, the root of a tree, the fall of a hillside. This one is oriented by a doctrine held by people in other countries too, and it is the first rule in the collection to arrive from outside the archipelago. This pack once called it the only one; the Madurese tanean lanjang, whose langgar closes its west end, makes that note wrong. What survives is the narrower claim: here a whole house plan is turned by it, there one small building is placed at the head of an arrangement. | |
| 1Moves no measurement | Canon | The number of treads is odd. Not roughly odd and not usually odd: odd, and the tradition says so. It is the only parity rule in this project, and it can fail in the way parity fails — by one. | |
| 1Moves no measurement | Canon | The number of bays is odd, and the house is named by it: rumoh lhee ruang, rumoh limong ruang. Odd because the middle bay has to be the middle — the raised room sits on it, and an even count would put a joint where the centre should be. | |
| 0Moves no measurement | Canon | Zero nails in the frame. The toi pass through mortises cut clean through the posts and are pegged and lashed — a frame that can move without breaking, on ground that moves. The Nias omo answers the same problem by triangulating every bay; this is a different answer and an equally good one: not stiff, but limber. | |
| 3Moves no measurement | Canon | Three parts across the width: the seuramoë keuë at the front for guests and the men, the raised tungai in the middle for sleeping and for birth, the seuramoë likôt behind for the women’s work. The sequence is not a staircase of rank but an order of how far in a person may come. |
Bibliography
dall-1982 · ethnography
Dall, G., The traditional Acehnese house, in J. Maxwell (ed.), The Malay-Islamic World of Sumatra (Monash University, Melbourne, 1982).
depdikbud-aceh · reference
Departemen Pendidikan dan Kebudayaan, Arsitektur Tradisional Daerah Istimewa Aceh (Proyek Inventarisasi dan Dokumentasi Kebudayaan Daerah, Jakarta).
waterson-1990 · ethnography
Waterson, R., The Living House: An Anthropology of Architecture in South-East Asia (Oxford University Press, Singapore, 1990).
Which of these matters most
Nearly every dimension on this page is the author’s own. What the provenance bar does not say is whether that matters. Each row below is computed by rebuilding the house: one dimension is pushed 20%, and the house is measured to see how far it follows. Nothing is estimated — the figures are the difference between two runs of the generator.
- 2.60 m
- 2.00 m
- 2.00 m
- 0.62 m
- 0.42 m
- 0.07 m
- 0.01 m
If only one dimension could be surveyed, this is the one. This is sensitivity, not uncertainty: it shows what moves, not how likely the value is to be wrong. The 20% is itself the author’s choice.
Invariants
There is no measured drawing yet, so correctness rests on structural truth. These checks fail the build.
- passBilateral symmetry about Symmetric about the north–south mid-plane, z = 0 — and deliberately not across its width
952 points; largest mirror deviation 0.00 mm.
- passEvery tenon contained in its mortise
42 pasak, takik and tumpu joints; every one engaged.
- passNo joint spans more than one stage
Every joint engages two parts from the same stage or from consecutive ones.
- passStage ordering, and nothing below ground
94 parts, every one bearing on something already standing or on the ground.
- passMesh integrity: indices, triangles, normals
200 triangles; indices in range, none degenerate, every normal unit length.
- passThe house is long east–west, and the ridge lies along it
13.0 m along the ridge against 7.1 m across it. The reason is prayer and this check cannot see it: what a model can hold is which axis is long. The other nineteen buildings here are turned by something local; this one by a doctrine held in other countries too.
- passThe ladder has an odd number of treads
9 treads, rising 0.29 m at a step to a floor 2.50 m up. The count is not declared but falls out of those two figures — which is exactly why the rule can be broken by a centimetre.
- passThree parts across the width, with the middle one raised
The front veranda → The raised middle room → The back veranda, with the tungai 0.35 m above its verandas. What the height states is not a guest’s standing but a boundary: how far in an outsider comes.
- passNo nails in the frame: threaded beams, pegs and lashings
42 joints, every one a threaded beam or a lashing. A frame that moves without breaking, on ground that moves — the Nias omo answers the same problem by triangulating every bay, and this is a different answer: not stiff, but limber.
- passThatch courses lap with no bare band
25 courses from eave to ridge.
- passEvery part declares the dimensions it came from
94 parts, every one tagged. By least-sourced input: 0 measured, 0 canon, 94 interpolated.
- skippedChecked against a measured drawing of a real house
Skipped. No measured drawing has been wired in, so there is nothing to compare against. This check will not be made to pass by weakening it.
What a check is preventing
Ten green rows are not a reason to believe anything. So one check is run against a house built to break it, and what is printed below is that check’s own verdict rather than a description of it.
Move the rise of a tread by a centimetre and the house does not move at all: the floor is at the same height, the ladder still reaches it, every step is still comfortable, the frame is untouched. All that changes is that the number of treads is now even, and the tradition says odd. It is the only counterexample in this project that turns on a count rather than a length — and the only one where the difference between the sound house and the broken one is a single piece of wood.
The house as built
Rear prow raised to 0.31 m
The second house is never rendered. It is built during the build, failed, and thrown away.
This bar moves as real surveys are wired in. That movement is the project's progress metric.