(Up to: Terrorism Is Communism Emergent Network Protocols Exploring An IDScheme )

Originally a post to the E-Mint mailing list, this seems like a relevant node to have. The original story refers to a statistical project being funded by the US National Science Foundation under its "Approaches to Combat Terrorism" program, but the implications for the research are somewhat wider than merely looking out for people planning to blow things up.

My post: :

On Tue, 12 Oct 2004, Caroline Teunissen wrote:
>> TROY, N.Y. - Amid the torrent of jabber in Internet chat rooms - flirting by
>> QTpie and BoogieBoy, arguments about politics and horror flicks - are
>> terrorists plotting their next move?

I don't get how studying *public* chatrooms would help uncover *private*
conversations. Terrorists aren't so stupid as to discuss their latest
plans in plain text on MSN Chat, surely?

As it's difficult enough to track down existing net bogeymen running huge
attacks against major websites, who probably *do* already use
distributed and possibly-encrypted private chatrooms, I don't quite get
how monitoring public rooms would help at all. I guess the tour de force
would be if you were intercepting traffic at the ISP level, and could
match up conversations based on the destination, but that's a different
quntity of data altogether...

Unfortunately I get the impression that there's a much more interesting
level to the research that's been obfuscated by the terrorism angle...

>> If, for instance, RatBoi and bowler1 consistently send messages within
>> seconds of each other in a crowded chat room, you could infer that they were
>> speaking to one another amid the "noise" of the chat room.
>>
>> "For us, the challenge is to be able to determine, without reading the
>> messages, who is talking to whom," Yener said.

This has possibly greater implications than being able to scan for
terrorists talking about their latest devious plot. My guess is that
interaction remains strong in a public arena between those familiar in a
private arena, and this is a way of inferring the latter from the
former. If it works, it allows monitoring parties to put together
reasonably-accurate social networking maps, i.e. who's "friendly" with
whomever else on the net. On a large enough scale, this would allow you to
"see" where clusters of social groups form, and the nature of the group
based on a few known key members.

In terms of how terrorists (and hackers, alike) "organise" themselves,
this could be an efficient tool to determine who to monitor further,
alongside traditional "face-to-face" investigational techniques. (e.g.
Arrest one suspect, and they're likely to give up names of peers readily.)

Naturally, the same method can be applied to any non-formal organisation,
for any reason [[1][4]].

Oh, and there could be some interesting commercial opportunities for such
a scheme, too. Link-finding seems to be all the rage these days, and an
emergent process to do so could be a crowd-puller...

.g

[1] <[http://uk.indymedia.org/en/2004/10/298741.html][5]>