A collaborative effort to generate a raw, but broadly functional, prototype of the Interoperable Private Attribution (IPA) proposal.
The goal of this project is to explore a stricter threat model than the original proposal assumes, with the potential for malicious but non-colluding servers. That is, servers (or helpers) that do no collude with each other, though they may collude with user agents (browsers or mobile operating systems), websites, or mobile apps.
The tools here will not include fully functional servers and clients, but instead command-line utilities that operate on files.
The focus is on readable code that is easy to understand and modify. This code might be useful for evaluating performance claims, but only in general terms; a full and performant implementation is likely to be much faster.
A token effort will be made to meet privacy and security goals. However, this is only for the purposes of learning. No serious effort will be made to ensure that these goals are met and therefore this code is not fit for production use.