4 ms·

Counterpoint: https://blog.pinboard.in/2011/11/the_social_graph_is_neither/ https://blog.pinboard.in/2011/11/the_social_graph_is_neither...

• jancsika 10y ago
  That counterpoint has a strong point here: > You can either have a decentralized, communally owned social graph (like Fitzpatrick envisioned) or good privacy controls, but not the two together. We already have a communally owned social graph in the form of publicly viewable FLOSS mailing list archives. If the lack of good privacy controls opens up the community to risks, then one should be able to assess the risks by mining the data in those archives. So has anyone ever written a paper about mining FLOSS mailing list archives for fun/profit?
  • Chris2048 10y ago
    > You can either have a decentralized ... social graph ... or good privacy controls, but not [both] Might homomorphic encryption change this?
    • jancsika 10y ago
      My point is that we don't have good privacy controls on FLOSS mailing lists now-- in terms the social graph it represents, the metadata, and the content. So the question is: what risks do FLOSS mailing list archives pose, as measured by analyzing the publicly viewable data in those archives? To answer your question: technically, yes. But I don't think it will practically change things until we have clear, concrete ways to measure how what we're already doing is bad.