Commit graph

104 commits

Author SHA1 Message Date
CRT
d8d12915da Give the contemporary kit the gate it never had, and fix the five models it immediately caught: ninety coplanar faces across the kit, a waste and a tap hole built as open tubes that culled to black, and a barrier end that two readers in a row took for a fault. 2026-09-02 07:57:41 +07:00
CRT
314d24cc64 Make the fold and squeeze sliders do something: both poses live in anchor rotations that no rebuild re-applies, so the controls were inert while the methods worked. 2026-09-02 07:27:59 +07:00
CRT
275a76d1f2 Add the lectern: a panelled shaft cut to the slope of the desk it carries, so the two meet along a plane rather than along one edge with a hand's width of air behind it. 2026-09-02 07:12:29 +07:00
CRT
f6ed95ad1c Add the A-frame pavement sign on a hinge whose origin is the pin, with a stay that collapses because both its links are driven rather than carried, and fix the support gate that reported it floating. 2026-09-02 06:55:55 +07:00
CRT
b305546964 Add the park bench, and with it the slat-with-gap rule: the pitch is derived from the run and the count so no array leaves a sliver at one end, on a cast frame whose arch is closed by its web rather than by fat legs. 2026-09-02 06:28:59 +07:00
CRT
3947246546 Add the picnic table: a plank array on one A-frame written once and built twice, with legs extruded from a parallelogram so they are sawn flat at both ends rather than standing on four edges. 2026-09-01 08:30:00 +07:00
CRT
107055ebb3 Add the concrete traffic barrier on the published section, and with it an extrusion for the constant-section family: a half-lap end at both ends, drainage cut by a second section rather than painted on, and a stepped run so the concrete has somewhere to vary. 2026-09-01 08:10:18 +07:00
CRT
0f5ddac950 Wire the last three contemporary models into the catalogue: they were built and committed without it, so nothing round-tripped them through the GLB gate and none appeared in the viewer. 2026-09-01 07:50:52 +07:00
CRT
3d0fdc0a9d Add the pedestal basin, and with it a plan sweep for the bodies a lathe cannot turn: a D plan with a flat back, per-ring depth and displacement, and the wall pinned down the whole height. 2026-09-01 07:50:51 +07:00
CRT
36e2cbc02d Add the fire extinguisher: the kit's first small-part cluster, under a lever whose travel is measured against the jaw it closes onto. 2026-09-01 07:17:26 +07:00
CRT
4c435861bb Add the gas cylinder: one revolve carrying a radial guard, which is the first array in the kit to stand on a lathe. 2026-09-01 06:46:27 +07:00
CRT
f8e7aed5df Give the vase a wall and a bore: a one-sided shell has no inside, so anything looking into it lost the far wall to back-face culling. 2026-09-01 06:37:35 +07:00
CRT
4f98247cf4 Add the blind critic, and take the pepper mill from 68 to 78 with it. 2026-09-01 06:33:24 +07:00
CRT
bb4896eca6 Show the contemporary kit in the inspector, by letting a catalogue entry belong to whichever registry wrote it. 2026-09-01 06:19:32 +07:00
CRT
d63d4911c6 Add the street bollard, and let a reference be shot for a model that does not exist yet. 2026-09-01 06:15:02 +07:00
CRT
e0a315a43f Round the mug's foot and its handle, and let an arc bar carry a section with more than four sides. 2026-09-01 06:09:33 +07:00
CRT
9ab88aa633 Rebuild the pepper mill as the tube it actually is: a steel collar, a recessed band of vertical flutes, a flat top and a knurled knob. 2026-09-01 06:01:43 +07:00
CRT
59a4559ab2 Take back the pepper mill rebuilt against a reference that asked for the wrong object, and record in the generator why a hint must not invent a material. 2026-09-01 05:54:11 +07:00
CRT
6a65991ec3 Serve either kit from the reference generator, and rebuild the pepper mill against the photograph it never had. 2026-09-01 05:49:23 +07:00
CRT
8f7ceb1abb Let the catalogue hand its overscroll back to the rail, so the blocks below it are reachable with the pointer over the list. 2026-09-01 05:37:47 +07:00
CRT
5bc455ee0c Give the catalogue its own scroll and move the tour and the recorder to the top, so the controls are not behind four screens of model names. 2026-09-01 05:24:02 +07:00
CRT
4a037ba816 Let the recorder capture what you are doing rather than only what the tour does, and letterbox the preview so the framing you compose is the framing you get. 2026-09-01 05:18:11 +07:00
CRT
fe2520d620 Put the kit's npm page, its source and its install line at the top of the inspector's rail. 2026-09-01 04:56:05 +07:00
CRT
5b051ca782 Keep the vegetable heap a heap past the slots it was authored for, give every repeated model its own seed, and stop the stalls' stock standing in each other. 2026-09-01 04:30:15 +07:00
CRT
fd9c26f3ba Add a page you can walk the square in, sharing the sun, sky and layout with the flythrough so the two cannot drift. 2026-09-01 04:17:10 +07:00
CRT
bcaadc144d Lay the market out like a market: four stalls with goods on their counters, stock stacked at their feet, and the camera walking the row instead of past it. 2026-09-01 04:05:16 +07:00
CRT
198d394e87 Put the square on real ground under a real sky, light the fires, and record upstream seeding the kit in its init template. 2026-09-01 03:52:58 +07:00
CRT
3722720d43 Fence the square with the kit's own rail and leave the houses out of the released cut, so everything in frame is either the kit or the ground it stands on. 2026-09-01 03:40:38 +07:00
CRT
801183ae25 Give the square a floor and a frontage, run the fences as boundaries, and route the camera down a lane instead of into the well. 2026-09-01 03:34:25 +07:00
CRT
d5c816320d Build the market square and a camera that walks through it, and cut the flythrough offline through the rasteriser rather than off a screen capture. 2026-09-01 03:15:29 +07:00
CRT
9b4ce16c59 Bring the anvil's horn back to its reference proportion, rebuild the book's metal furniture so no two overlapping faces share a plane, and release 0.2.0. 2026-09-01 03:00:37 +07:00
CRT
3f57165d06 Install the vibe-model authoring skill and record the state of both kits in PROGRESS.md. 2026-08-28 20:16:39 +07:00
CRT
af53b5ee92 Add the first five contemporary props, and give the contact shadow a per-pixel height falloff. 2026-08-28 20:06:59 +07:00
CRT
91847838ed Shade with vertex normals in the renderer, and add crease-angle normal smoothing so turned objects stop banding. 2026-08-28 19:37:07 +07:00
CRT
02c0711442 Fork the kit into contemporary-props with a nineteen slot palette, a per-model check that needs no registry round trip, and the first model. 2026-08-28 19:03:50 +07:00
CRT
77b926312b Aim the render camera at the picture's centre, include the contact shadow in the fit, and shoot the kit scene from the front. 2026-08-27 21:22:14 +07:00
CRT
b4aeccd2a2 Fit the render camera to the geometry and repack the kit scene, then regenerate the images. 2026-08-27 20:52:32 +07:00
CRT
5d793c83d0 Remove em dashes from README.md, REFERENCE.md and the registry README. 2026-08-27 20:18:43 +07:00
CRT
be8c7a485d Bring the repo's own front page up to what actually exists now
The root README was written when none of this was published, and it showed.

The install block was BROKEN in the same way the registry README's was, and I
fixed that one without looking here:

    bunx vibe3d init
    bunx vibe3d add @medieval-kit/wooden-barrel

`init` seeds `models.json` with `@scifi-kit` only, so the second line stops with
`Registry @medieval-kit is not configured in models.json`. The registry entry
and the two dependencies are now in the block, in the order they have to be run.

The npm link pointed at a file in this repository rather than at the package,
because when it was written there was no package to point at. There is now.

Neither document linked the live viewer at all — the one thing a stranger
landing here can use without installing anything. It is the second thing on the
page now.

REFERENCE's "publish your own registry, you need nobody's permission" section
gets the receipt: this registry is live and vibe3d was not involved. It also
gets the warning I would have wanted before doing it — npm stamps the
publishing account's email into the packument per version and later account
changes do not rewrite it, so the address on the account at that moment is
public permanently.

Checked every outbound link in both documents as a logged-out stranger: bun.sh,
vibe-stack/vibe3d and medieval.crt.fyi all 200. npmjs.com answers 403 to a
scripted request, which is its bot protection and not a broken link — the
registry API serves 0.1.1 as latest.

Still 404 for a stranger, and deliberately: github.com/CRTkafa/medieval-kit and
its /issues, which npm renders as the Repository and Issues links on the
package page. They come back the moment the repository goes public.
2026-08-27 17:28:17 +07:00
CRT
f2a8e304ce 0.1.1: drop baseUrl from the setup the README tells people to copy
Found by installing 0.1.0 from npm into an empty directory and doing exactly
what the README says, which is the first time anyone has done that — until an
hour ago there was nothing on npm to install.

Everything held except one line. `bunx vibe3d init`, the `models.json` registry
entry, `vibe3d add @medieval-kit` (46 files), `three` and `@types/three`: all
correct. `@types/three` is genuinely not optional, confirmed by removing it and
watching three TS7016s land on the runtime files init had just written.

The Project setup block was wrong. I wrote it by copying this repository's own
`tsconfig.json`, which carries `"baseUrl": "."` — and this repository pins
typescript at ^5.9.2, so it never sees that a fresh `bun add -d typescript`
installs 7.0.2, where the option has been REMOVED:

    error TS5102: Option 'baseUrl' has been removed.

So the one instruction on the page that a reader cannot check against anything
else was the one that stopped them. `paths` resolves relative to the tsconfig
on its own and has since 4.1, so the block just loses the line; the same config
now typechecks the installed source clean on 7.0.2.

The lesson is the one this repo keeps relearning in a new costume: a setting
copied from a working project is evidence about THAT project. This one was
right here and wrong in the only place it was published to.

Verified after the change in the same clean project: tsc silent, and
`createModel()` from the npm-installed hand-cart builds 2044 triangles with its
floor at y = -0.0000 — so the wheels that started this whole thread are on the
ground in the published package too.
2026-08-27 17:22:54 +07:00
CRT
909bfa0f9f Land the mill's ladder, and stop the table calling six models static
Two things, both found by pulling on the handcart's thread.

THE LADDER. I said three models had parts floating a few millimetres. One did.
The other two readings were my own measurement error and worth writing down,
because it is the same error as the vertex-sampling ones: I measured each part
against the model's BOUNDING BOX minimum, and a part deliberately bedded into
the ground drags that datum below the floor, so every correctly placed part
beside it appears to float by exactly the bedding depth. The forge's base
stones settle 6.8 mm and the well's drystone kerb 9.3 mm, both on purpose, and
the bellows and the well frame sit exactly on the floor their own source
declares. Nothing wrong with either.

The post mill's tail ladder was real: 14.5 mm clear. It already carried two
analytic corrections for the same corner, both correct about what they
describe, and neither accounting for the chamfer — the corner they compute the
drop of has been cut off, so the residual was very nearly the chamfer itself.
Rather than add a third derivation to a spot that has now been got wrong three
times, the assembled ladder is measured and put on the floor. A bounding box
cannot be wrong about this the way arithmetic can, and it stays right when the
chamfer or the rail section changes.

`restsOn` now covers the mill's trestle and ladder too, and its doc comment
carries the bedding caveat, so the next person to reach for it knows the datum
is the bounding box and knows what that costs.

THE ANIMATED COLUMN. `catalog-table.ts` decided it with `built.action`, which
is the VIEWER's hand-registered button and exists for five models. So the
generated table published the grindstone, the stone well, the post mill, the
cauldron, the forge and the handcart as static — six of the eleven models that
carry actions, including three of the most animated things in the kit.

The obvious replacement is worse. `core/kit.ts` wraps update as
`update: (dt) => options.update?.(dt, runtime)`, so it is a function on every
model whether or not that model does anything with it, and switching to it
marked all thirty-seven animated. One flag too narrow, its replacement too
broad.

So the generator now runs each model — a second of frames, then its action if
it has one — and asks whether anything moved. That catches what moves by
itself. What moves ON DEMAND it cannot catch, because the motion is behind an
interface the catalogue never exposed, so `Entry` now carries `actionNames`
from the model's own actions. Eleven models, and both documents list them.

Worth keeping: this table is generated precisely so it cannot go stale, it says
so in its own prose, and it was wrong for months. Generated is not the same as
correct — it only moves the mistake into the generator, where nobody looks.

Also in the npm README, which is the page npmjs.com renders: `vibe3d init` is
now the first line rather than unmentioned, `@types/three` is called out (three
ships no declarations, so without it the installed source does not typecheck),
the `@/` alias and `allowImportingTsExtensions` have a Project setup block, and
the one relative link is gone — npm resolves those against the repository, and
this one pointed at a directory the package deliberately does not ship.
`homepage` now points at the live site instead of a repository that is private.
2026-08-27 16:36:24 +07:00
CRT
bc78f295af Size the viewer to the viewport it has, not the largest one it might get
Reported as: in fullscreen the left rail cannot be scrolled to the bottom, it
looks clipped.

I could not reproduce it. In this environment the browser pane will not
composite frames, so there is no real wheel input and no screenshot, and
`requestFullscreen` needs a gesture and is refused — so the fullscreen case
specifically is untested by me. Everything I could measure came back healthy at
1920x1080, 900x620 and 600x900: the rail scrolls, reaches its last element,
has 17 px under the final control, carries no overlay across its bottom, has no
nested scroll container stealing the wheel, and `100vh` equals `innerHeight`.

What follows is therefore hardening against the mechanism that produces exactly
this symptom, not a confirmed fix, and it is written that way in the CSS.

`#app` was `height: 100vh`. `vh` is the LARGEST the viewport gets, not the size
it currently is, and where the two differ the grid stands taller than the
window. The body clips its overflow, so that difference is not scrollable, it
is gone — and what goes is the bottom of the rail, the one part of this page
with more content than fits. `100dvh` tracks the viewport actually in use, with
`100vh` left above it for anything that does not know the unit. In an ordinary
window they are the same number, which is why this can only ever appear
somewhere they come apart.

`min-height: 0` on the rail closes the neighbouring hole: a grid item's
automatic minimum size is its content, and the browser relaxes that to zero
only for as long as the item stays a scroll container. It is one line of CSS
away from not being one, and then the row grows past the window and takes its
own scrollbar out of reach.

Verified after the change at all three sizes: the grid matches the window
exactly, the body never scrolls, the rail still reaches its last element, and
the stacked narrow layout still splits 414 / 486.
2026-08-27 15:49:14 +07:00
CRT
7bc879b223 Remove a scratch directory that came in with the last commit
`my-registry/undefined/` — ten files, a copy of core and one barrel. It is the
output of a `vibe3d add` run made from a directory with no `models.json`
resolved against it, so the CLI wrote the install under the string
"undefined". It was a pre-publish audit's scratch work, not part of the kit,
and `git add -A` swept it up.

Nothing referenced it and nothing shipped it: the npm tarball is driven by
package.json "files" (dist, README.md, LICENSE) so it was never going to reach
the registry. Removed all the same, because a directory called "undefined" in a
repository someone is meant to read from is its own kind of defect.
2026-08-27 15:43:15 +07:00
CRT
11e20a7b1a Put the handcart's wheels back on the ground
Reported as "the wheels are in the wrong place, it is floating", and it was
three separate faults in one wheel.

The one that showed: the part declares `origin` at the axle so `setRoll` turns
the wheels in place, and a part with an origin has its geometry written
RELATIVE to that point — the anchor supplies the rest. The geometry was ALSO
translated by the axle height and position, so both were applied twice. The
wheels sat 304 mm off the ground at twice the axle height, and the cart stood
on its two shaft tips.

The one hiding behind it: the felloe segments were turned into place and then
moved out to the rim. That looks equivalent to moving out and then turning and
it is not — `rotateZ(a)` sends local +X to (cos a, sin a) and local +Y to
(-sin a, cos a), so a piece placed at (cos a, sin a) has its ARC pointing
radially. Every rim segment was a stubby block sticking 89 mm past its own
felloe, straight through the iron tyre meant to bind it. `cart-wheel` has had
the correct order all along, with a comment saying why; this model did not
follow it.

And the hub: a lathe is built around Y, the wheel is drawn in the XY plane, so
the hub stood upright inside the wheel like a little drum instead of running
along the axle. Invisible behind the spokes, which is how it lasted.

Then a fourth, found while checking the fix. The ground meets a pole's SURFACE,
not its axis: the shafts ran their axis to y = 0, so the square-cut end of an
inclined pole buried its low corner 13.7 mm, and the cart rested on two buried
tips with its wheels still fractionally clear. The axis now stops one tip
radius short and wheels and shafts share one floor.

`restsOn` is the check that would have caught all of it. The support check asks
whether every piece is connected down to the ground, and a cart standing on its
shaft tips answers yes: bed hangs off shafts, wheels hang off bed, nothing
floats. It reported 0 failing cases throughout. Connectivity cannot tell a
wheel from a decoration, and the renderer cannot either — the contact shadow is
cast from the model's own lowest point, so the shadow looked right the whole
time.

What the geometry cannot know, the author does. Naming the parts that carry the
weight turns "it rests on its wheels" into arithmetic, at 1 mm, which is tight
enough to have caught the 13.7 mm as well as the 304 mm. Declared on the cart,
the trestle table, the bench and the market stall.
2026-08-27 15:42:49 +07:00
CRT
a038f622dd Put the config step before the command it is required by
`vibe3d add @medieval-kit` resolves the namespace through `models.json` and
stops dead without it:

    Registry @medieval-kit is not configured in models.json

The registry README printed the two install commands FIRST and the
`models.json` entry underneath, introduced with the word "first". A reader
copies the top block, runs it, gets that error, and leaves. `vibe3d init`
seeds only `@scifi-kit`, so this is every new reader, not an edge case.
Config block first now, commands after it.

The cover card said `bunx vibe3d add @medieval-kit/wooden-chest`. Both address
forms are valid, but the bare namespace resolves through the registry's
`defaultItem` and installs the whole kit, which is what someone looking at a
card of the whole kit is being offered. A card has room for a command and not
for the `models.json` entry that command needs, so the shorter promise is the
honest one — and naming one arbitrary model implied a precision the card cannot
deliver.

Verified against the CLI rather than inferred from it:

    add @medieval-kit/registry   -> Registry @medieval-kit has no item named registry
    add @medieval-kit            -> Would install @medieval-kit · 46 files
    add @medieval-kit/grindstone -> Would install @medieval-kit/grindstone · 10 files
2026-08-27 15:26:45 +07:00
CRT
b29762a540 Draw the card with a command instead of a browser
The card a shared link unfurls to was a picture of the models with no title on
it, and the models on it were too small to tell apart. Both are fixed, but the
part worth writing down is how.

The first attempt was an HTML page composed in the site's own fonts and
screenshotted. It looked right and it was wrong: the model count on it becomes
a picture of a number the moment the shutter closes, and `check-docs.ts`
cannot read a picture. That is the exact failure that check exists for — the
READMEs claimed 27 models when there were 37 — and a card is the worse place
to have it, because nobody reading one can see the kit to know it is stale.

So `bun run cover:build` counts the catalogue and draws the numbers it counted.
`site:build` runs it, which means the published card cannot be older than the
models it shows, and the day a model is added the card says 38 without anyone
remembering to make it.

Drawing type needed type. `scripts/text.ts` reads the outlines out of a `.ttf`
and fills them with a scanline pass — four sub-scanlines a row with exact
horizontal coverage, which is more code than sampling a 4x4 grid and visibly
better at the 16px the two mono lines are set in. It is deliberately partial:
format 4 cmap, simple and composite outlines, advances, no kerning or shaping.
What it cannot draw it throws on, because a headline missing a letter is not
something you notice before it is published. Archivo and IBM Plex Mono come
with it under the OFL, the same two faces the site is set in.

The renderer moved to `scripts/raster.ts` so the card and the contact sheet
draw from the same code and cannot disagree about what a barrel looks like;
`render.ts` is now the command line over it, unchanged in what it does. Its
one new flag is `--ground`, a flat background instead of the per-cell
gradient: forty gradients side by side read as banding, and on the page's own
ink the models float on the card as one field instead of sitting on it as a
grey slab.

Ten columns by four, because 1.91:1 and 37 models want 40 cells and at that
shape the cells come out square, which is what lets a barrel look like a barrel
at 116 pixels. The last row is centred — three empty cells left flush is a
notch in the corner, and a short row centred is how a contact sheet ends.

Byte-identical across runs, so it does not churn the diff on every build.
2026-08-27 15:00:15 +07:00
CRT
dfee59381e Give the published page a card
A link to the site showed a URL and a blank square. It now carries a title, a
description, a canonical URL, an icon and a full set of OpenGraph and Twitter
tags, so a page nobody has opened yet arrives as something you can look at.

Two of those tags have to be ABSOLUTE and it is not a style question: a
scraper fetching the page has no base to resolve `/cover.png` against and
drops it silently. `build-site.ts` now copies the images into the output as
well, because an `og:image` pointing at this host has to have the file ON this
host — a card that 404s shows nothing, and nothing about the page itself looks
wrong while it happens.

The cover is a real 1200x630 render rather than a crop of a square one, which
took a change to the renderer worth having anyway. `newFrame` and
`frameCamera` both took a width and a height already; only `renderOne` was
insisting they be equal. And with a non-square frame the old camera fit showed
its flaw: it fitted the bounding SPHERE, which is one number and wrong for
anything not roughly round. The kit scene is wide and flat, so its sphere is
far taller than the scene is, and fitting that into the vertical field left the
models in a band across the middle with empty air above and below — most of a
1.91:1 card. It now fits the BOX, iteratively, because perspective is not
linear in distance: place the camera, project the eight corners, move by how
far the worst one lands past the edge. Two passes is inside a pixel.

That change improves every render, not just the card: on the contact sheet
every model is visibly larger in its cell and nothing crops.

`check-docs` now watches `viewer.html` too. Its description quotes the model
count and the triangle total, which is the same sentence that went stale three
times in the READMEs before that check existed — and a stale number in a card
is worse than one in a README, because nobody reading the card can see the kit
to know it is wrong. The description says "37" rather than "thirty-seven" for
exactly that reason: a word is not a number the check can find.

Also repointing the npm README's two images from `raw.githubusercontent.com`
to the live site. They were pointing into a private repository and would have
404'd on npmjs.com; now they do not depend on the repository being readable at
all.
2026-08-27 14:34:59 +07:00
CRT
151aeb89b2 Take the oak out
Five attempts and it was never good. The last one is better than the cabbage
that prompted the complaint, and "better than the version you called ugly" is
not a standard worth shipping against — every other model here would have to
carry it.

The constraint underneath is real and worth recording, because it is not a
tuning problem and no sixth attempt was going to find it. The kit's budget is
2500 triangles a model. The oak wanted a bole you can see, limbs radiating out
of it, and a dome of foliage over the lot; at 2500 that is roughly 128 for the
trunk, 880 for the boughs and 1400 for the crown, and 1400 buys about forty
masses. Forty masses hung on twelve limbs is twelve clusters, and twelve
clusters do not make a dome — from a third of the compass you see the gaps
between them. A turntable says so plainly where a single three-quarter view
does not, which is the other thing to keep: I judged this model on one angle
for four rounds.

Everything else in this kit is something a person made, with flat faces and
straight edges, and lowpoly is not a compromise for those — it is close to how
they are actually built. A tree is the one subject where the budget genuinely
bites, and the honest options were to except it from the budget, to abandon
realism for a deliberately stylised shape, or to drop it. Dropped.

Out with it goes the `leaf` slot and its palette key, since the oak was the
only thing using either. Back to thirteen slots. `bark` stays, because the log
pile uses it; `limestone` stays, because the trough and the grindstone do.
`scripts/probe-tree.ts` goes too — it measured one model against one pair of
photographs and there is no longer a model.

37 models, 30 390 triangles, thirteen material slots, all checks pass, 0
floating cases, GLB round-trip clean, Conformant, documentation in step.
2026-08-27 14:25:26 +07:00
CRT
c02ec0dcbe Put the documents back in step
The previous commit changed the tree's triangle count and I pushed it with all
three documents still stating the old total. `check:docs` caught it — it exited
1 inside the same command that ran the commit, and the commit went through
anyway because it was not what the shell was waiting on.

Which is the check working and the process around it not. The check belongs
where a failure stops something: it is in CI, so the push is already flagged,
but a gate I run beside a commit rather than before it is a gate I can walk
past. Worth writing down rather than quietly fixing, since the whole reason
`check:docs` exists is that these numbers had gone stale three times before
anyone noticed.
2026-08-27 14:07:40 +07:00
CRT
e99969aad9 Rebuild the oak's crown, and stop asking it to be one blob
"Very ugly" was the report and it was right. Put beside its reference the
diagnosis is not subtle: the crown read as crumpled paper, the colour was flat,
and the trunk had been swallowed whole.

The correction that mattered was to a rule I had invented myself. I decided
early that "a crown is one piece" and made the probe fail anything else, so
every version since had been tuned until the flood fill returned 1 — which is
the instruction "be a single smooth blob", and that is exactly what it kept
producing. The support rule requires the crown to be connected to the TREE, not
to itself: a clump that swallows a branch end is held up by that branch
whatever its neighbours do. Letting the masses separate is what buys the gaps,
and the gaps are the silhouette. The probe now reports the piece count as
information instead of grading it.

Four other things, each of which cost a render to see:

  - The clumps were three rings on five sides — a coarse bipyramid — and
    sixty-five of them creased every facet against its neighbour. Five on eight
    fixed the surface and made a cabbage. Four on six is the middle, and the
    middle is where it belongs.
  - `roughenGeometry` came off the crown entirely. Ragged edges read as foliage
    where smooth ones read as stone: true of a SILHOUETTE, false of a surface.
    All it did was crumple the skin. Raggedness has to come from the
    arrangement.
  - The tone between clumps varied by +/-0.09, so every mass was the same green
    and the crown was one object with facet shading on it. Widening it to
    +/-0.19 put seven whole clumps through the floor into black, because the
    tint jitter, the mottle and the baked occlusion all pull the same way
    underneath. It is asymmetric now, -0.03 to +0.23, which is also truer: a
    lit mass is far brighter than the base green while a shaded one is filled
    in by the sky.
  - Every clump sat on the outer half of its limb, making a crown that is a
    RING — full at the rim, hollow in the middle, and from a third of the
    compass you look straight through the hole. I had that right once and
    removed it while cutting triangles, which is the cheapest regression to
    make and the hardest to notice: nothing failed, the model just got worse.

The bole went from 0.037 of the height to 0.046. The reference's own number
came off a photograph of a tree seen whole from a distance; at this size, under
a crown, 0.037 disappears and what is left is a cabbage on a stick.

Twelve limbs instead of nine, spread more evenly, with cheaper geometry paying
for them. And the turntable is now the thing I judge it by rather than a single
three-quarter view — one good angle hid two that were full of daylight.

Honest state: better than it was and not yet good. Three of six angles read as
a tree; the others are still bunched. 38 models, 32 798 triangles, all checks
pass, 0 floating cases.
2026-08-27 14:06:58 +07:00