Lisp to Python translation criticism?

[Mark McEahern]

> TMDA is no doubt very effective at preventing spam, but I see several
> problems with it:
> - it requires a manual response from everyone who wants to send you email,
> at least for the first mail, or some pretty sophisticated configuration

I speak from ignorance, but can't you prepopulate your whitelist?

> - what happens if someone you've pre-approved mails you from a
> temporary/transient/borrowed address?

They'll either respond to the confirmation or not--assuming they don't, you
can still see them in the unapproved queue, can't you?  I think your point
is that's a false positive.

> - what about legitimate email that doesn't have a human at the other end
> (e.g., Web forms, mailing lists, announcements, "spam proofed" email
> addresses etc.)?

Isn't that what the dated email addresses are for?  Namely, I buy a book at
amazon---oops, surely I'm not going to change my email for amazon every time
I buy something from them.  But I guess I'd already have added amazon's
address (or its entire domain)?

> It seems to me that the potential for false positives is much higher using
> TDMA than w/ a content filtering technique.

I'm still curious to try both and see, in practice, which one hits the sweet
spot for me.

Thanks for your thoughtful reply.

Cheers,

// mark

-