Brian Tillman said:
The strings in rules are not case-sensitive. But "v1agra" and
"viagra" and "vîagra" and "viagrá" and "vi-a-gra"are not the same
thing.
Nor will "via<i></i>gra" in an HTML formatted e-mail get detected. In
forums, I sometimes have to insert a null set of tags (they produce no
effect) to get around some stupid spell checker. I'll enter the word
"embarassment" and their ignorant swear word checker that doesn't know
how to parse words will make it "embar***ment". That's no so bad in
this case but in some cases it obliterates some or all of the word. So
if the forum support HTML then I simply insert an HTML tag pair to slice
up the word at the bad spot. Spammers use it to hide words from word
parsers because I haven't found one that handles HTML in Outlook
[Express], other e-mail clients, or anti-spam software.
Any product trying to eradicate or detect "bad" words should be
processing itself against both the HTML code (which is all text) and
also against the rendered version. But then, as Brian mentions, using
v-ia-gra, v1agra, and other forms bypasses the word list filters, so
obviously you need a better approach to eradicating spam. If the OP
wants to know what are good anti-spam products, best would be to ask in
the alt.spam group.