[IMC-Tech] Captcha implementation: which way to choose?

[MJ Ray]

Mike Tonks wrote:
> It does say it passes content & IP Addresses to a blogspam central
> server for testing purposes, but I guess that could be disabled for
> Indymedia.

I don't see how to do cooperative/distributed bayesian content
analysis without sending the content to a server running the analysis
routine.

Two levels of protection appear possible:

  1. someone could run a blogspam.net central server for IMCs that
doesn't use IP Addresses in its decision-making (disable those
plugins);

  2. IMCs don't send the IP Addresses to the central server.

I suspect not sending the IP Addresses to a server which *does* use
them (like blogspam.net itself) will result in a higher false-positive
rate, but I don't know how much higher.

Hope that explains,
-- 
MJ Ray (slef) Webmaster and developer for hire at | software
www.software.coop http://mjr.towers.org.uk        |  .... co
IMO only: see http://mjr.towers.org.uk/email.html |  .... op