Multi-party aggregation proposal using oblivious bucketization

[erik-anderson]

Microsoft has been looking at potential improved protocols for the aggregate measurement multi-party compute approach currently described in the Aggregate Reporting API proposal.

We've published a document for how the oblivious bucketization might work.

We'd love feedback on in what ways it's perceived to be better or worse than existing proposals. We'd also like to give an overview in the scheduled calls we're having for measurement.

Feedback in this issue on via issues in the other WICG repo the doc is published in would be appreciated.