Graham's spam filter

[Edward K. Ream]

> This technique is evil, and fails to scale in any useful way.

> I know that if I'm emailing someone who uses one of these systems, I won't
> bother. I have too much to do as it is.

Sorry, but these statements are completely wrong, for several reasons:

1. I have a right to filter who can email me. There is nothing evil about
asserting that right.  Think of ChoiceMail as asking _strangers_ to
introduce themselves.  ChoiceMail is less restrictive than, say, having an
unlisted telephone number...

Yes, some _friends_ may be left off the original whitelist, and that is the
only drawback I know of ChoiceMail; it is slightly inconvenient for them to
"register".  But it is very easy for them to do so: I've gotten "no reason"
and "because" as reasons.  That's all they had to type!  Because I knew
them, that was sufficient reason to accept them.  We are talking about a
one-time cost of less than a minute.  Real strangers should not, I think,
object to telling in a few words why I should let them talk to me.

Important note: there are separate lists for people who haven't filled out
the form people who have filled out the form.  The latter list is almost
always empty, so it hard to miss the few legitimate replies that do come
along.

2. As I said in my original posting, we can all think of reasons why the
ChoiceMail trick might not work exactly for newsgroups.  But that's not very
interesting!

I'd like to elaborate on this point.  My late friend and mentor, Prof. Bob
Fitzwater, once explained how he does research.  The first two steps are
becoming familiar with the literature.  One reads articles looking for a)
techniques that work, and are immediately useful for the problem at hand and
b) techniques that don't immediately work, but might be interesting.  We all
know that creativity is largely a subconscious process; the more "slightly
relevant" information we can feed the mind the more likely we are likely to
make creative leaps.  It was in this spirit that I mentioned ChoiceMail; it
may not be completely relevant to the task at hand, and it does solve a
related problem supremely well.

Edward
--------------------------------------------------------------------
Edward K. Ream   email:  edream at tds.net
Leo: Literate Editor with Outlines
Leo: http://personalpages.tds.net/~edream/front.html
--------------------------------------------------------------------