On Mon, Jun 15, 2009 at 15:42, Kurt J<[email protected]> wrote: >> Not sure what the relevance of crypto is. > > i suppose mixed meanings here. i thought your concern was taking > anonymous data and tracing it back to users by reversing a hash or > something like this.
I wasn't thinking of that particular scenario, but I would be surprised if someone has naively claimed eg that hashed usernames are anonymous, which would be silly given dictionary attacks. >> http://www.eff.org/deeplinks/2009/02/last-fm-and-the-diabolical-power-of-data-mining > > right. > > so are you envisioning a scenario where the RIAA uses anonymous data - > which only contains listening habits and a hash code - to uniquely > match you with your myspace profile? and uses this to sue in court? > despite what the above blogpost says data record matching is a tricky > problem. i think this is a paranoid scenario. I'm not envisioning any particular scenario, and I agree that many are paranoid. I'm just very uneasy with claims that anonymization is easy or that users should be given high expectations of anonymity from supposedly anonymized data. > but, you have a valid position. and maybe i'm in the minority. altho > you did agree to openly share your personal listening habits earlier > :-) There's no although. I both don't want to make guarantees of anonymity for others and do want to my non-anonymous data made as widely available as possible. > i still believe that any obscure risk posed by anonymous user data is > out weighed by potential benefits. I believe that the risks posed by non-"anonymous" user data is outweighed by the benefits of raw and complete data. Mike _______________________________________________ Libre-fm mailing list [email protected] http://lists.autonomo.us/mailman/listinfo/libre-fm
