>>Shelby wrote: >> But as I said in previous post I know of at least one proprietary way to detect >> this. > > Valdis wrote: >Do tell. Apply to this message. I will alert you when the patent is applied for and has a link online at www.uspto.gov. I hope that will be within a few months, but the problem I have right now is getting enough non-spam data to test the algorithm outside of the theoretical. SpamArchive.org gives me spam data. But I need non-spam data also to form a non-theoretical corpus test. Can any one help? >Dean wrote: >No, it isn't. And it is an illegal method, because you (if you are an >ISP), probably don't have permission to block non-spam mail. Conceptually agree. Concept of attempting to kill spam by stopping the relaying of all unauthenticated email when email protocols and installed base are not designed that way is like burning a forest because you want to lower the mosquito population. You kill a lot of trees and other things that the forest was designed for, and you may or may not kill a lot of mosquitos. Worse, the mosquitos can reproduce faster than the trees and other things destroyed. If you change the protocol and understand the tradeoffs of the change, then that could be okay: http://www1.ietf.org/mail-archive/ietf/Current/msg22215.html That is why I proposed a way to move non-spam bulk email (by new definition) out of email. Obviously that was not popular here at IETF and for reasons frankly that I can understand and agree with to a certain extent (although I think the alternative is worse as pointed out some where burried in that thread). That was an attempt to see if it is possible to make services such as DCC legal, because otherwise they can be misused to block non-spam bulk email. Note afaik DCC is most often implemented at ISP MTA level (per Dean's point about US federal communications law). (Note I did not say that DCC is not useful, if not misused and understand it can be subverted in future). And to legalize the blocking of bulk email at other nodes in the channel. Now I can __gleefully__ see it isn't likely to happen (gleeful that my proposal failed). Which is fabulous for the value of the anti-spam algorithm I invented recently. My point is (see link in previous posts) that all existing anti-spam in public domain that I am aware of is of dubious benefit, legality, and most have the potential (eventually if the "mega spam virus" occurs) to be as harmful as they can benefit. People are not "kooks" for working to improve a state-of-the-art which is currently all dubious hueristics. There may or may not be "kooky" people working on anti-spam (I don't care), but their methods can't be any more "kooky" than the current *published* state-of-the-art: http://www1.ietf.org/mail-archive/ietf/Current/msg22184.html I just hope that the world of people using anti-spam doesn't have to suffer a horrific experience or experiences (similar to Harald's discovery which lead to this thread, but on larger scale) to become aware of the dangers of existing tools. Thanks, Shelby Moore http://AntiViotic.com