Adam Shostack writes:
Timothy C. May wrote:
| I don't think "escrow" is the salient feature of what Adam is hinting at, | or at least it's now what I'm focussing on here. Rather, I think the really | intriguing thing is using _logical_ names for remailers, and not | necessarily tying them to specific accounts, specific account owners, or | specific sites. ("Call by name," if you will.) This could make remailers | more persistent.
Netescrow is a system that Matt Blaze proposed to do key escrow such that the recovery of a key must be public.
Someone wrote me a while back asking about using oblivious key escrow ("netescrow") to build a remailer. I was a bit skeptical of using exactly the mechanism that I outlined in my paper on the subject, but I think the idea raises some intersting avenues to look at. I'll try to dig up the message I sent on the subject. -matt The paper in question, by the way, can be found at: ftp://research.att.com/dist/mab/netescrow.ps