Re: [PLUG] OT: Spam

[gabriel rosenkoetter <gr@eclipsed.net>]

On Wed, Oct 22, 2003 at 06:46:01PM -0400, Kevin Brosius wrote:
> gr wrote:
> > He said exactly what I would about this. In summary, negative
> > results are meaningless in this case. Positive results are
> > meaningful.
> Maybe, if you take a very strict view.  But I don't look at it as one
> big spammer.  So yeah, some will de-munge addresses.  Other's will not. 
> Depends on what you think the ratio is.

But that's precisely why the results of the study you cited are
meaningless!

> Seems to me that spam filtering is in the same arms race.  I don't think
> this argument applies just to de-munging.  (So, yes, I agree with you
> about the arms race.  I just don't see how you can apply against one of
> the methods and not the other.)

I don't argue that spam filtering isn't part of the same arms race,
but the spammers have no knowledge of whether my filtering worked on
their messages or whether I simply deleted them by hand. So they
have no way of adjusting their tactics based on my behavior. They
have every opportunity to adjust their tactics based on address
munging in a publicly-accessible archive precisely because it is
publicly accessible.

> Welllll, I can think of at least one flaw in that study that would prove
> otherwise.  But with the little information available on the page I
> couldn't prove or disprove it.

"I have a great proof for this theorem, but it won't fit in the
margin."

It's unfair to say "there's a problem with that study" and expect us
to believe you without providing at least a short description of
what you think the problem is.

Is it so serious as trying to prove something that it is logically
impossible to prove?

> You've failed to help me understand how the munging makes web archives
> less "easily readable".

I can only say the same things so many times, Kevin, but I'll say
them one more:

- it makes it difficult to contact the original author
- it damages (as in "makes non-functional") a historical record
  with regard to validation that two copies of that record match and
  with regard to cryptographic signatures
- it doesn't work anyway

> Have you used MARC[1]?

Once. Hated the interface, and thought it was a bit duplicitous.
(Supposing that mailing list didn't WANT to be archived publicly,
what then?)

What is less readable about their munging?

On Wed, Oct 22, 2003 at 07:16:24PM -0400, Douglas wrote:
> What if a script were to replace all of the email addresses in all of the
> messages that were sent too this list with links to a form where you could
> send that email address a message directly (to ask that they send their
> email address to you).

That is unacceptable.

- It's still harder for a reader to contact the original sender. In
  the twenty-years version, whose to say that the Perl script or
  whatever behind the emailing process will still work? All you can
  rely on is non-volatile bits. Munging destroys those bits.
- It still damages the historical record in exactly the same ways.
- It also (probably) won't work. If you provide a web form to send
  email, a web spider can make use of it. If you rate-limit the email
  sent through that web form, you may very well inconvenience a
  legitimate user.

-- 
gabriel rosenkoetter
gr@eclipsed.net

Attachment: pgptRS6dNZYLg.pgp
Description: PGP signature