111 reputation
3
bio website surfaceeffect.com
location Ipswich, United Kingdom
age 40
visits member for 1 year, 11 months
seen Jan 13 at 14:40

I'm a software architect, mostly for web based systems on Unix, and an interaction designer (although I prefer the term interaction architect, since I'm more about efficiency and getting the job done than making things pretty).

I'm now running my own consulting firm focussed primarily on interaction design, but I do seem to get my hands dirty in code very frequently!


Sep
18
awarded  Teacher
Sep
14
comment Is there a digital “safety deposit box” equivalent?
As for this not being equivalent to an independent third party, that depends entirely on who the keyholders are. I suggested that some of them should be uninterested third parties (such as a lawyer). But having interested parties involved too protects against fraud by, or compromise of, the "trusted" third party.
Sep
14
comment Is there a digital “safety deposit box” equivalent?
Actually, you'd need more than two keyholders, since as you've written it if either Bob or Charlie loses their keys you're stuck. So you'd need Daniel to have key 1 & 3. Now any one of the three can lose their keys and you can still get the data.
Sep
14
comment Is there a digital “safety deposit box” equivalent?
ah, ok, yes, hence the use of 3DES rather than 2DES. So you would have to consider your algorithm selection with some care - good point.
Sep
14
comment Is there a digital “safety deposit box” equivalent?
No, I think you misunderstood, I wasn't suggesting you give the only key to many people. I was suggesting that you encrypt in (say) three layers, and then give each key to (say) two people. So if one of the keyholders dies before Alice, you have a backup for that layer. But because there are two other layers the data is still safe. Perhaps I should redraft to make that clearer.
Sep
14
comment Is there a digital “safety deposit box” equivalent?
I thought I'd addressed the issue of all the keyholders being needed by suggesting that each key could be known to more than one person. Could you elaborate on the risks of encrypting something which is already encrypted. I understand you'd need entirely unrelated keys, but I don't see what the risk is so long as that is the case. Notably Truecrypt allows you to encrypt using multiple algorithms with independent keys - although I presume that is to protect against defects being found in an algorithm in the future. What have I missed? I assumed than none of the parties can be entirely trusted.
Sep
6
answered Is there a digital “safety deposit box” equivalent?
Sep
6
awarded  Autobiographer
Sep
6
awarded  Supporter