Community and membership architecture

A group that grows past the size where everyone knows everyone needs some way to say who belongs, who has earned what standing, and how decisions get made. The ordinary answer is a database holding everybody's details, which makes the organisation the single most attractive object in the room to anyone who wants that list.

There is another way to build it. Cryptographic membership proofs let a person demonstrate that they belong to a group, and that they hold whatever standing the group grants, while which member they are stays with them alone and the group's records stay clear of it. The proof carries the property; the identity stays with the person.

I build these — the membership layer, the governance shape that sits on top of it, and a straight account of what the architecture protects and where its protection ends. That last part is the part that matters, because much of what gets sold as privacy technology promises more than its design delivers.

The honest limit, first: this works at scale and gets thin below it. A proof of membership among six people identifies you nearly as well as your name would, and the anonymity becomes real somewhere north of fifteen or twenty. I say that before taking the work.

Once a group grows past the size where everyone knows everyone, it needs a way to say who is a member, who has earned what, and how decisions get made. The usual answer is a database with everyone's details in it — which turns the organisation into the most valuable thing in the room for anyone who wants that list.

There is another way. Certain kinds of maths let a person prove they are a member of a group, and prove what standing they hold, while which member they are stays with them alone and the group's records stay clear of it. The proof carries the fact; who you are stays with you.

I build these: the membership part, the decision-making part on top of it, and a straight account of what it protects and where that protection ends. That last part matters most, because much of what gets sold as privacy technology promises more than its design delivers.

The limit, said first: this works at scale and gets thin below it. Proving you are one of six people points at you nearly as well as your name would, and it starts working properly somewhere past fifteen or twenty. I say so before taking the work.