I should have asked this years ago when there was more traffic/users on this list (and I was using a different nym[1]) ... but years ago there was not such established proof of data correlation in "anonymized" data sets (netflix, "anonymized" medical data, etc.) So, in the post-netflix, post-aol-user-data[2] world, I am doing the following to sanitize my online nyms: - establishing a distinct user account, on a different provider, for each topic I discuss. For instance, this nym only discusses encryption/"yro"/tor subjects. - Scrubbing my language of certain idiosyncratic phraseology or blatant aspects of "style" ... favored slang, etc. Perhaps even purposefully bad grammar in finer points of grammar that most people get wrong - so as to keep the "correct" forms from being identifiable. - Forged user agent / email agent headers - Never referencing my own _actual_ published works or comments, or linking to myself or my work product in even a second-hand manner. These actions are probably not robust. My thought is that it can lower my common profiles below the noise threshold that a particular aggregator might have. What do _you_ do to sanitize/de-correlate your nyms ? [1] obnymleakage [2] http://en.wikipedia.org/wiki/AOL_search_data_scandal