Implementing telemetry well is hard. On the one hand, some projects would like to know as much as possible about their users so they can focus their development efforts where they would be the most effective. On the other hand users are rightly reluctant to give up what is their private data. Data can range from configurations and use of features to performance.

Numerous projects have tried, some more successfully than others. The Linux Foundation's Telemetry and Data Collection Policy provides one example of a framework for projects to operate within. A recent example of the tensions can be seen in the discussions following a proposal to introduce telemetry into the Fedora Desktop.

Meanwhile various sophisticated techniques are being developed that try to square the circle of obtaining detailed telemetry while simultaneously preserving privacy and making it hard for people to deliberately skew the data. One example is the IETF's Distributed Aggregation Protocol and another is Google's RAPPOR.

This session is an opportunity for projects interested in these developments to share experiences and consider what might be gained by collaboration.