Comment How can they publish it ? (Score 1) 326
After reading this article I was quite disappointed.
It's an extension of whitelist mechanism with some graph theory included. ( and quite too much theory for something so simple...)
As stated in the article:
It sounds like BCC: and there is nothing special about it. It's quite widely used. Most of my spam comes without CC: or multi recipients.
It would certainly be very interesting for ISP, because they can track many emails at a time. The main issue would be the size of the graph induced (less CPU intensive -> more MEM intensive) and privacy.
But it's definitely not for individuals.
I keep this for generating examples for bayesian analysis, that's all:
Also nice graphs... perhaps to big at the end...
It's an extension of whitelist mechanism with some graph theory included. ( and quite too much theory for something so simple...)
As stated in the article:
The most obvious contermeasure is to never use multiple recipients in the To: or Cc: headers of a spam message.
It sounds like BCC: and there is nothing special about it. It's quite widely used. Most of my spam comes without CC: or multi recipients.
It would certainly be very interesting for ISP, because they can track many emails at a time. The main issue would be the size of the graph induced (less CPU intensive -> more MEM intensive) and privacy.
But it's definitely not for individuals.
I keep this for generating examples for bayesian analysis, that's all:
Using this algorithm, it would be possible to generate training sets for learning algorithms in achieving accurate, automated spam filtering.
Also nice graphs... perhaps to big at the end...
