On Tuesday, April 2, 2002, at 11:11 AM, dmolnar wrote:
>
>
> On Tue, 2 Apr 2002, Tim May wrote:
>
>> Imagine N transponders. Coded sequences are broadcast, recipients are
>> unknown. (Actually, _everyone_ receives, but only some can decode.)
>
> Sounds vaguely like the setting for this paper:
>
> Xor-Trees for Efficient Anonymous Multicast and Reception
> Shlomi Dolev, Rafail Ostrovsky
> http://citeseer.nj.nec.com/dolev98xortree.html
>
> Abstract:
> In this work we examine the problem of efficient anonymous broadcast and
> reception in general communication networks. We show an algorithm which
> achieves anonymous communication with O(1) amortized communication
> complexity on each link and low computational complexity. In contrast,
> all
> previous solutions require polynomial (in the size of the network and
> security parameter) amortized communication complexity.
>
Yeah, but O(1) is easy to achieve for "small" systems (I'll leave it to
others to draw parallels to the buildability of Bernstein's machine).
Of course, having "only" N nodes, with N being as small as 50, would
still be useful.
BTW, the cypherspace interpretation of of transmitters and receivers is
not new. Gelernter's LINDA tuple space acts this way. It's in my 1988
Manifesto, and we talked about it at length 1989-91. I gave some talks
at Hackers and called it "Democracy Wall," a la China and the grafitti
deposited untraceably (modulo physical observers, but within the domain
it was moderately untraceable) and read untraceably.
Usenet with ramailer-to-posting and universal reading acts this way,
which is why my instantiation of Blacknet used Usenet.
It's fundamentally a blackboard system, a la AI, with a shared space
into which one can deposit and retrieve.
Scaling is over-rated, in my admittedly contrarian view. Scaling and
limits are great for theoreticians to worry about and write papers
about, but most things don't get done because they don't get written and
built, not because they don't scale well.
(The "flash crowd" effect seen with napsters is a big problem, but is
more related to the lack of ontologically-sound throttling mechanisms,
e.g., pricing mechanisms. The local record store would be mobbed ("it
doesn't scale") if CDs were being given away for free. The problem is
not with topology, but with pricing, the allocation of scarce resources.)
--Tim May
--Tim May
--
Timothy C. May [EMAIL PROTECTED] Corralitos, California
Political: Co-founder Cypherpunks/crypto anarchy/Cyphernomicon
Technical: physics/soft errors/Smalltalk/Squeak/agents/games/Go
Personal: b.1951/UCSB/Intel '74-'86/retired/investor/motorcycles/guns
Recent interests: category theory, toposes, algebraic topology
