| Random Page
HardSecurity is necessary to break concerted attacks: you must meet force with force. -- SunirShah
- I've been thinking a little about this lately. I wonder what the possibility is that we might meet force in kind. I.e., rather than performing admin-level blocking, thwart attackers' handiwork using the same techniques they use. To whatever extent possible, the goal would be to avoid the sort of hard-line authoritarian approach of HardSecurity.
- The best means of achieving this would likely be to create an autonomous AI agent to monitor the wiki. We could program or teach it to recognize potential community threats and respond by restoring pages to their prior contents. Somewhere it would keep a log of all it did so that whoever operated it could overrule it later.
- If an attacker chooses to use a set of scripts to post spam, and does so in a distributed fashion, there is little that HardSecurity can do to prevent it (apart from something like ZWiki:ShieldsUp). However, if a script can be used to automate posting spam, then an intelligent script can be used to automate deleting spam.
- I guess there's no reason a priori that an adaptive or fuzzy approach couldn't be built into HardSecurity instead of imposed from the outside. I think, however, that an intelligent agent is a more elegant solution. It avoids the atmosphere of an authoritarian environment and lessens the attraction of circumventing HardSecurity. I think it might also slightly lessen the frustration of legitimate posts being denied. At least they'd remain up for a few minutes. ;-) (We'd have to make a good effort to collectively determine the agent's behavior so that the community could feel that they contributed to it, rather than having it unilaterally impose some ad hoc standard on them.)
- I'd like to develop this idea further, if anyone cares to comment. Maybe start putting together a basic decision tree or something of the sort. I'm not sure as yet exactly how I'd go about implementing it, but it's a good candidate for my long-term to-do list. (Read: somewhere in the year-or-two range.) If everyone's interested enough, we could move discussion to a separate page. -- anon.
This sounds like a ChannelBot on InternetRelayChat...
meet force with force -- see IteratedPrisonersDilemma for good reasons not to go soft.