Analyzing Privacy Aspects of the W3C Vibration API

I didn't consider the de-anonymisation tactic of marking or communicating to an observer, that's a new tactic I hadn't considered.

I was expecting a de-anonymisation scheme along the lines of fingerprinting how the device vibrates given a certain pattern. Because there's a human element - the DeviceOrientation readings will be impacted by the user holding the device - you may be able to pick up the user's rough age/gender/strength/weight by judging how much their body alters that fingerprint.