Sources — Rumah gadang — Rumah Adat Nusantara

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	Citation
Length of one ruang along the ridge axis.ruangLength	3.10 mIf 20% out: 3.10 m	Interpolated	No source
Depth of one lanjar, front to rear.lanjarDepth	2.50 mIf 20% out: 1.50 m	Interpolated	No source
Number of lanjar. The sources name the lanjar as a division of the interior but do not fix a count for the ordinary house, so this number is the author’s.lanjarCount	3If 20% out: 2.50 m	Interpolated	No source
Section of a tonggak, the underfloor post.postSection	0.20 mMoves no measurement	Interpolated	No source
Height of the batu sandi a post stands on.padHeight	0.25 mIf 20% out: 0.05 m	Interpolated	No source
Width of the batu sandi.padDiameter	0.46 mMoves no measurement	Interpolated	No source
Clear underfloor height, from the ground to the floor frame.kolongHeight	1.90 mIf 20% out: 0.38 m	Interpolated	No source
Height of the wall at its foot, from the deck to the wall plate.wallHeight	2.10 mIf 20% out: 0.42 m	Interpolated	No source
Outward lean of the wall. The sources state that the body widens toward the plate; none gives the angle, so this figure is the author’s.wallLean	8°If 20% out: 0.06 m	Interpolated	No source
Depth of a rasuak, the floor-frame beam.floorFrameDepth	0.26 mIf 20% out: 0.05 m	Interpolated	No source
Thickness of a floor board.deckThickness	0.05 mIf 20% out: 0.01 m	Interpolated	No source
Width of one floor board.deckBoardWidth	0.26 mMoves no measurement	Interpolated	No source
Thickness of a wall board and of a bilik partition.wallThickness	0.06 mMoves no measurement	Interpolated	No source
How far the anjuang floor steps up above the main floor. That the step exists is canon; how high it is, is not.anjuangRise	0.42 mIf 20% out: 0.08 m	Interpolated	No source
Height of the ridge above the wall plate, at mid-span.ridgeRise	3.20 mIf 20% out: 0.64 m	Interpolated	No source
How far the ridge line sags at mid-span relative to its ends.ridgeSag	0.42 mIf 20% out: 0.08 m	Interpolated	No source
Rise of both ends of the ridge. Symmetric: unlike the tongkonan, neither end is the higher one.ridgeEndRise	0.50 mIf 20% out: 0.10 m	Interpolated	No source
How far the roof projects beyond the end of the body. This is where the gonjong lives, so the projection is deep: the roof edge has already left the house before it starts to climb, which is why the long façade still reads as a straight eave.ridgeOverhang	2.60 mMoves no measurement	Interpolated	No source
How far the ridge curve overshoots its control point before rising to the end; this is what makes the ridge a curve rather than a kink.ridgeUpsweep	0.20Moves no measurement	Interpolated	No source
Height of a gonjong tip above the end of the ridge. The tips stand well above the ridge end, and the hollow between a pair of them is the most recognisable part of the silhouette.gonjongRise	2.60 mMoves no measurement	Interpolated	No source
Where a gonjong tip stands across the roof, as a share of its half-width. A pair stands wide — footed at the roof edge and leaning inward toward the tips — so the separation is a property of how wide the roof is rather than a length of its own. The author’s reading of the form; no source gives the figure.gonjongSplay	0.58Moves no measurement	Interpolated	No source
How far the lift reaches back into the body, as a share of its half-length. Nearly all the curve happens outside the house; it encroaches slightly so the transition is not a crease on the gable line. A large value lifts the edge too early, so the roof has flattened to ridge height before it reaches the end wall and there is no gable left.gonjongReach	0.12Moves no measurement	Interpolated	No source
How late the roof edge turns upward into the gonjong. A value of one lifts it in a straight line from the end of the body; a high value holds the eave level and then raises it sharply near the tip.gonjongSweep	2.20Moves no measurement	Interpolated	No source
Section of the verge member at the tip relative to its foot. This is what makes the point a point rather than a cut-off stub.gonjongSparTaper	0.25Moves no measurement	Interpolated	No source
Radius of the verge member that carries the gonjong out to its point.gonjongSparRadius	0.06 mMoves no measurement	Interpolated	No source
How far the eave oversails the outer post line, so the drip falls clear.eaveOversail	1.35 mIf 20% out: 0.27 m	Interpolated	No source
How far the eave hangs below the wall plate.eaveDrop	0.90 mIf 20% out: 0.18 m	Interpolated	No source
Depth of the wall plate the rafters bear on.plateDepth	0.24 mMoves no measurement	Interpolated	No source
Width of the wall plate.plateWidth	0.16 mMoves no measurement	Interpolated	No source
Width of a rafter in section.rafterWidth	0.07 mMoves no measurement	Interpolated	No source
Depth of a rafter in section.rafterDepth	0.11 mMoves no measurement	Interpolated	No source
Number of rafters per ruang, per side.raftersPerRuang	3Moves no measurement	Interpolated	No source
Rafters added beyond the per-ruang count, for the two projecting ends.raftersAtEnds	4Moves no measurement	Interpolated	No source
Radius of the ridge beam.ridgeBeamRadius	0.08 mMoves no measurement	Interpolated	No source
Radius of a bamboo purlin.purlinRadius	0.04 mMoves no measurement	Interpolated	No source
Position of the upper purlin, as a share of the run from ridge to the roof break.purlinAboveKnee	0.50Moves no measurement	Interpolated	No source
Position of the lower purlin, as a share of the run from the break to the eave.purlinBelowKnee	0.55Moves no measurement	Interpolated	No source
How far the roof boarding sits above the rafter axis, as a share of rafter depth.sheathingOffset	0.60Moves no measurement	Interpolated	No source
Clearance between the roof boarding and the first ijuk course.ijukBedClearance	0.02 mMoves no measurement	Interpolated	No source
Exposed depth of one ijuk course.ijukCourseDepth	0.32 mMoves no measurement	Interpolated	No source
How far a course stands proud of the one below it.ijukThickness	0.10 mMoves no measurement	Interpolated	No source
The share of a course that the course above laps over.ijukLap	0.40Moves no measurement	Interpolated	No source
Thickness of the singok, the triangular board closing the gable end.singokThickness	0.05 mMoves no measurement	Interpolated	No source
Width of a rasuak, as a share of the post section.sillWidth	0.90Moves no measurement	Interpolated	No source
Depth of the dish in the batu sandi that the post foot seats into, as a share of stone height.postSeat	0.30Moves no measurement	Interpolated	No source
How far a post head runs up into the floor frame, as a share of frame depth.tenonRun	0.70Moves no measurement	Interpolated	No source
Depth of the tested joint engagement, as a share of the smaller member.jointEngagement	0.30Moves no measurement	Interpolated	No source
The front faces the halaman, with the rangkiang ranged across it. The rule is relational rather than a compass bearing — and this is where the rumah gadang differs from the tongkonan, which faces north.orientation	0°Moves no measurement	Canon	Navis, A. A., Alam Terkembang Jadi Guru: Adat dan Kebudayaan Minangkabau (Grafiti Pers, Jakarta, 1984).
The ruang count is always odd: three, five, seven, nine.ruangIsOdd	1Moves no measurement	Canon	Navis, A. A., Alam Terkembang Jadi Guru: Adat dan Kebudayaan Minangkabau (Grafiti Pers, Jakarta, 1984).
Under the Koto Piliang laras the floor at both ends is raised into anjuang; under Bodi Caniago the floor is one level plane. The difference in adat is legible as a difference in floor height.anjuangKotoPiliang	1.00Moves no measurement	Canon	Navis, A. A., Alam Terkembang Jadi Guru: Adat dan Kebudayaan Minangkabau (Grafiti Pers, Jakarta, 1984).
The base form carries four gonjong: one at each corner of the roof, where the eave lifts into a point.gonjongBase	4Moves no measurement	Canon	Departemen Pendidikan dan Kebudayaan, Arsitektur Tradisional Daerah Sumatera Barat (Proyek Inventarisasi dan Dokumentasi Kebudayaan Daerah, Jakarta).
A bilik is added as each daughter marries; their number is a record of how the matriline has grown.bilikAreTally	1Moves no measurement	Canon	Waterson, R., The Living House: An Anthropology of Architecture in South-East Asia (Oxford University Press, Singapore, 1990).
The bilik fill sequentially from one end with no gaps. That they are added one at a time is canon; the direction of fill and the absence of gaps are the author’s, and they are why this house is not symmetric in its partitions.bilikFillOrder	1.00Moves no measurement	Interpolated	No source
The body walls lean outward, widening toward the wall plate.wallsLeanOut	1.00Moves no measurement	Canon	Schefold, R., Domenig, G. & Nas, P. (eds), Indonesian Houses Vol. 1: Tradition and Transformation in Vernacular Architecture (KITLV Press, Leiden, 2003).
Pegged joints; the frame goes up without nails.noNails	1.00Moves no measurement	Canon	Schefold, R., Domenig, G. & Nas, P. (eds), Indonesian Houses Vol. 1: Tradition and Transformation in Vernacular Architecture (KITLV Press, Leiden, 2003).
The posts stand on the grid of ruang and lanjar, symmetric about the transverse mid-plane.postsOnGrid	1Moves no measurement	Canon	Departemen Pendidikan dan Kebudayaan, Arsitektur Tradisional Daerah Sumatera Barat (Proyek Inventarisasi dan Dokumentasi Kebudayaan Daerah, Jakarta).
The rangkiang stand across the halaman facing the rumah gadang. This house’s orientation constraint is relational rather than compass-bound — and these are what it is relative to, so without them the rule says nothing.rangkiangOpposite	1Moves no measurement	Canon	Navis, A. A., Alam Terkembang Jadi Guru: Adat dan Kebudayaan Minangkabau (Grafiti Pers, Jakarta, 1984).
Width of the halaman between the house and the rangkiang.halamanDepth	10.00 mMoves no measurement	Interpolated	No source
Plan side of one rangkiang. Only the footprint is drawn: the rangkiang are not modelled, and that is already stated as an absence.rangkiangPlan	2.60 mMoves no measurement	Interpolated	No source
Height of the rangkiang floor above the ground. Only its massing is drawn; every figure here is the author’s.rangkiangFloorY	1.90 mMoves no measurement	Interpolated	No source
Height of the rangkiang body.rangkiangBodyHeight	1.60 mMoves no measurement	Interpolated	No source
Rise of the rangkiang roof above its body. A rangkiang’s roof carries gonjong like the house; this massing is a straight gable and states none.rangkiangRoofRise	3.40 mMoves no measurement	Interpolated	No source
Side of a rangkiang post.rangkiangPostWidth	0.26 mMoves no measurement	Interpolated	No source
Overhang of the rangkiang roof beyond its body, on the massing drawn here.rangkiangEave	1.00 mMoves no measurement	Interpolated	No source
Spacing between rangkiang along the halaman.rangkiangSpacing	5.00 mMoves no measurement	Interpolated	No source

Bibliography

navis-1984 · ethnography
Navis, A. A., Alam Terkembang Jadi Guru: Adat dan Kebudayaan Minangkabau (Grafiti Pers, Jakarta, 1984).
vellinga-2004 · ethnography
Vellinga, M., Constituting Unity and Difference: Vernacular Architecture in a Minangkabau Village (KITLV Press, Leiden, 2004).
schefold-2003 · reference
Schefold, R., Domenig, G. & Nas, P. (eds), Indonesian Houses Vol. 1: Tradition and Transformation in Vernacular Architecture (KITLV Press, Leiden, 2003).
waterson-1990 · ethnography
Waterson, R., The Living House: An Anthropology of Architecture in South-East Asia (Oxford University Press, Singapore, 1990).
depdikbud-sumbar · reference
Departemen Pendidikan dan Kebudayaan, Arsitektur Tradisional Daerah Sumatera Barat (Proyek Inventarisasi dan Dokumentasi Kebudayaan Daerah, Jakarta).

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.

1ruangLength3.10 mbody length
2lanjarCount2.50 mbody depth
3lanjarDepth1.50 mbody depth
4ridgeRise0.64 mridge height
5wallHeight0.42 meave height
6kolongHeight0.38 mridge height
7eaveOversail0.27 mroof reach
8eaveDrop0.18 meave height

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.

pass
Bilateral symmetry about the transverse mid-plane
22967 points (257 of 263 parts); largest mirror deviation 0.00 mm.
pass
Every tenon contained in its mortise
70 pasak, takik and tumpu joints; every one engaged.
pass
No joint spans more than one stage
Every joint engages two parts from the same stage or from consecutive ones.
pass
Stage ordering, and nothing below ground
263 parts, every one bearing on something already standing or on the ground.
pass
Mesh integrity: indices, triangles, normals
32272 triangles; indices in range, none degenerate, every normal unit length.
pass
The ruang count is odd
5 ruang; 6 ruang boundaries.
pass
The ridge sags in the interior; both ends rise to the same height
lowest point at s=0.50 (7.34 m); the two ends at 8.26 m and 8.26 m.
pass
Koto Piliang: the floor at both ends rises into anjuang
main floor 2.46 m, anjuang floor 2.88 m; a step of 0.42 m over 24 boards.
pass
The gonjong are part of the roof, not masts standing on it
4 gonjong; highest tip 10.86 m against a ridge end of 8.26 m; the roof reaches every tip within 0.07 m.
pass
The roof edge holds its line, then climbs past the ridge into the gonjong
level at 3.66 m as far as z=6.82 m, then rising 7.20 m over 3.53 m — a slope of 2.0 : 1 — to 10.86 m, past the ridge end at 8.26 m, narrowing to 3.13 m.
pass
Nothing in the roof frame reaches outside the roof
78 rafters and gable panels cut to the roof section they sit at; the furthest reaches 35 mm outboard and 54 mm low.
pass
The bilik fill the rear lanjar in sequence, with no gaps
3 bilik out of 3 interior ruang; consecutive, one ruang wide each.
pass
The walls lean outward; the body widens as it rises
wall foot 3.75 m from the axis, head 4.05 m; the outermost board reaches 4.07 m at 8°.
pass
Ijuk courses lap with no bare strip; the ridge is covered
35 courses; smallest lap 2.3% of the slope run.
pass
The eave oversails the outer post line
eave 5.40 m from the axis; outer post face 3.75 m; oversail 1.65 m.
pass
The eave clears the wall plate
roof at x=4.05 m sits at 4.68 m; top of the wall plate 4.68 m; eave 5.40 m.
pass
Post count follows the ruang and lanjar counts
5 ruang × 3 lanjar → 6 × 4 = 24 posts, 24 pad stones.
pass
Every part declares the dimensions it came from
263 parts, every one tagged. By least-sourced input: 0 measured, 0 canon, 263 interpolated.
skipped
Checked 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.

This check enforces the one claim in this pack that a reader standing in the room could disprove with a spirit level: under the Koto Piliang laras, the floor at both ends stands above the middle. Flatten the step and the house stops being able to say which adat raised it. The threshold is searched rather than picked — and the step stops being a step before it reaches zero, at the point where it is thinner than the board that forms it.

The house as built

pass2.46 m / 2.88 mmain floor · anjuang floor

main floor 2.46 m, anjuang floor 2.88 m; a step of 0.42 m over 24 boards.

Rear prow raised to 0.04 m

fail2.46 m / 2.50 mmain floor · anjuang floor

No anjuang was built. The step of 0.042 m is thinner than the 0.050 m floor board that would form it, so there is nothing to stand on.

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.