On Wed, May 01, 2002 at 12:13:53AM -0700, Kayvan A. Sylvan wrote: > On Wed, May 01, 2002 at 09:08:40AM +0200, Lars Gullik Bjønnes wrote: > > Mate Wierdl <[EMAIL PROTECTED]> writes: > > > > | So then removing attachments is just snakeoil: viruses are > > | application/octet-stream. > > > > Doeas all virii use octet-stream? I think not. > > > > | As for size restriction: it is again snakeoil with the proposed limits: > > | Klez is 120K. > > > > Did you do this only for the mis-benefit of klez? > > > > btw. filtering is the only way to be somewhat sure. > > Yes. Filtering is the only sure way.
Yes, here is how filtering works 1 I get a spam 2 I create a receipe which will block this spam 3 goto 1 In particular, filtering is retroactive. Please note that about two percent of all spam addressed to the lyx lists get actually sent to list members. Two percent, unfortunately, means at least one a day---would be tolerable perhaps, but some are viruses. And in 6 months, 2% may mean 10 a day. > > Excerpts from my procmail recipes (the last one kills klez): > Please show me the receipe that will stop tomorrow's new spam written in Spanish, Chinese, Korean, Turkish. Seriously, where do you guys get these laws "Filtering is the only sure way" and "filtering is the only way to be somewhat sure" ? Presently the only "almost" sure thing is that the sender of a spam/virus will not reply to the proper address of a confirmation request. > :0 > * Content-Type: (text/html)|(multipart/alternative)|(multipart/mixed) > * B ?? TVqQAAMAAAAEA > /dev/null > Can you think of a legitimate message that will be stopped by this receipe? How many spams do you recommend for me to read/analyze each day, and how many receipes do you recommend for me to create each day? Mate
