News for hackers. Stuff that matters.
Re:A Tale of Security Gone Wrong by Jtsummers on Thursday April 07, 2016 @06:21PM (#11451165)
The easy passwords (for today) were already filtered out. The idea was to future-proof it by recording how hard they were. Rounding to, say, whole numbers still leaks information, but a bit less. So now everything from [13.5,14.5) map to 14, that's still smaller than [0,+inf]. You could use some sort of hidden value mapping them to grades that aren't obviously mappings of entropy (like A, B, C, D). It indicates that there's something about them, but without seeing a table mapping [10,12] -> C you don't know exactly what. That's still a point for data to leak, though, should the mapping be discovered (encoded elsewhere in your database or somewhere in your code).
Re:A Tale of Security Gone Wrong by gavingmiller on Thursday April 07, 2016 @06:56PM (#11451362)
Your solution of mapping to a grading system A,B,C,D; or terrible, crap, better, best was mentioned in another thread about this article. It's a common thought, however it's incorrect because you're still leaking substantial information about passwords. By storing entropy of any kind: whole number, graded, > threshold, etc you are weakening your password hash. This is completely unnecessary where better solutions exist: TFA
Re:A Tale of Security Gone Wrong by Jtsummers on Friday April 08, 2016 @09:58AM (#11454671)
You've repeated what I wrote and I'm not sure why. I was responding to the original comment asking if rounding or similar could work and said that it still leaked data. Which is what you've written here in response to me.