Declared data
Data asserted by an actor about themselves or their products. Publishable as-is because it commits its author, but never presented as verified: telling it apart from observed data is what stops the platform from laundering a claim by republishing it.
Not to be confused with
- Derived data
Data the platform infers from real exchanges — for instance a point of sale inferred from a delivery passport followed by a receipt. Publishing an inference exposes two parties who did not consent jointly, which is why it stays subject to double consent.
What the blog says
- Who Sorts, When No One Proves Anything?
Joseph Stiglitz (Nobel 2001), screening, and why food sovereignty begins with verification: when the seller doesn’t signal, it’s the buyer who organizes the sorting.
- Il se peut que le progrès soit le développement d’une erreur
En 1962, Cocteau adresse une phrase à l’an 2000. Prise au sérieux avec Haber-Bosch : une réussite peut être le début du problème suivant.
- Orphée et le regard en arrière
Le mythe d’Orphée interdit de vérifier ce qu’on croit savoir. Une lecture de cette règle — via Cocteau et un capteur de température — pour distinguer une donnée d’une preuve.
- What Modernization Left Out
A century multiplied yields and divided description. What changed is not morality but cost: green it, or bring it back in.
- The Farming Debate Isn’t Too Complex. It’s Badly Described.
Latour and Spinoza: agricultural complexity isn’t excess chatter around a simple object — it’s an outdated apparatus of description. What actually changed is the cost of the singular.
- Du Normalbaum au label : pouvoir déplier la simplification
Un label rend une réalité complexe lisible, et c’est une fonction essentielle. Reste à pouvoir remonter derrière la catégorie — sans la supprimer ni la discréditer.
- Conforme. Puis : montre-moi.
Rabelais fait purger Gargantua avant de l’instruire. Ce que ce détail dit d’un système de conformité qu’on croit corriger en ajoutant des champs.
- Le catalogue des livres qui n’existent pas
Rabelais énumère 217 jeux, puis catalogue une bibliothèque d’ouvrages imaginaires. Ce qu’un registre trop nourri finit par prouver — et de qui il parle.
- What Our Registry Will Never Be Able to Write
“You can’t put everything on a form”: with James C. Scott, mètis, and the Prussian forest — what a registry can attest to, and what it must never grade.
- Our registry started too late
"You sell traceability and you cannot prove who writes your articles." The objection is valid, and what it reveals is worse than what it targets.
- Opacity Is Not a Lie
What a seventeenth-century philosopher has to do with traceability: why the opacity of an origin is not fraud but mutilated knowledge, to be restored rather than punished.
- The Stamp and the Thing
Ceremony or substance: with Spinoza, why a declaration of conformity is a rite — and how proof finally shows what the standard truly wants.
- Import Is Not a Sin
Import = bad, local = good? With Spinoza: the good is not a property but a relation. Traceability is a matter of proof, not morality.
- Four Hundred Thousand Sensors That Did Not Exist
eFishery reported 400,000 connected feeders at the edge of its ponds. There were 24,000. What its collapse says to anyone who sells traceability.
Where it is implemented
lib/discovery/types.ts