CLCTN · TYPE MANIFEST · ARTWORK · SEPTEMBER 2026
ART · THE WALL
One of one. Every assumption in the app breaks here.
Fifteen types identify an object by finding its twin in a catalogue. A painting has no twin, no model number, no barcode and no release date — the object in front of you is the only one that has ever existed, and proving what it is is the hobby. Photography then does something stranger: it reintroduces editions, so one type has to hold both the unique and the numbered.
You asked painting or photo. They are one type. A painting and a silver gelatin print share fifteen of eighteen fields, hang on the same wall, insure on the same schedule and fade under the same light — the real axis is unique versus editioned, and that is one field, not two types.
01 · FIVE THINGS ART BREAKS
{{ b.b }}
WHAT IT NEEDS
{{ b.needs }}
02 · ONE TYPE, TWO SHAPES
Switch the work in Tweaks. Same manifest, same add form, different fields awake — because editioned is a field, not a fork. A vintage print and a later print from the same negative differ by ten times in price and by one date field, which is the single most valuable thing this type can teach someone.
Note what the attribution line does as you switch. Certainty about who made it is a spectrum in this hobby, and every other type in the app treats identity as a fact.
{{ workName }}
{{ workSub }}
{{ workDims }}
{{ r.k }}
{{ r.v }}
{{ r.note }}
ATTRIBUTION · HOW SURE ARE WE
{{ attrName }}
{{ attrNote }}
THE LIGHT CLOCK · DERIVED, NEVER STORED
{{ lightHead }}
{{ lightNote }}
SAFE
{{ lightState }}
IRREVERSIBLE
Third type to want a clock, after wine and retro computers — and the first where the damage is silent and total. Nobody notices fading; they notice it in the photo from ten years ago.
WHY NOT TWO TYPES
Splitting would duplicate provenance, exhibition history, framing, dimensions, insurance and conservation — six hard features, twice, forever. And a collector who owns eleven paintings and four photographs wants one wall, not two shelves. Medium is a field; the app should not make it a schism.
03 · ARTWORK.YML
RESOURCES/TYPES/ARTWORK.YML
TYPE 17
{{ manifestYaml }}
04 · IDENTIFICATION — THE BACK, NOT THE FRONT
CAPTURE STRATEGY
Photograph the verso first.
The front is the part everyone photographs and the part that identifies nothing. The back carries the evidence: gallery labels, exhibition stickers, auction lot stencils, a stamped estate mark, the framer's ticket, a title in pencil on the stretcher. Same structural move as the retro computer's underside label — the identity is printed on the side nobody displays.
So the add flow asks for two photos, front and back, and runs OCR on the second one. Every string it finds becomes a suggested provenance entry the collector accepts or ignores — never an auto-filled field, because a gallery label is a claim rather than a fact.
No provider, no corpus, no barcode — and unlike every other type, that is permanent rather than a gap we might fill. What ships instead is an artist-name authority list so that Ansel Adams, A. Adams and Adams, Ansel stop being three artists on one wall.
05 · WHERE THE MONEY CHANGES THE APP
A shelf of records is worth four thousand euros and a wall of art can be worth four hundred thousand. Nothing about the interface should shout about that, but three features stop being optional.
{{ m.k }}
{{ m.t }}
{{ m.b }}
06 · STILL OPEN