News for hackers. Stuff that matters.
Re:Why SpiderOak doesn't de-duplicate data across users by rarrrrrr on Friday August 27, 2010 @09:09PM (#1640863)
... and yet SHA1 collisions have happened. According to NIST, "Federal agencies should stop using SHA-1 for...applications that require collision resistance as soon as practical"

Regrettably, we don't actually have any truly ideal hash functions yet.

Re:Why SpiderOak doesn't de-duplicate data across users by tptacek on Friday August 27, 2010 @10:48PM (#1641022)
Not that any of the rest of this argument makes much sense, but SpiderOak (says it) uses SHA256. There are no induced collisions on the horizon for SHA256, and there won't be accidental ones in this data set.
Re:Why SpiderOak doesn't de-duplicate data across users by klodolph on Friday August 27, 2010 @09:26PM (#1640897)
I did a Google search for a SHA-1 collision and I thought nobody had found a collision yet, they had just found a collision attack in the 2^50-2^60 range.

Of course, turning a collision into an exploit in this case would be challenging. (Preimage attacks ⊂ collision attacks.)