Re: OT: writing resumes with VT100 for a Lisp job

From: Thomas Schilling (tjs_ng_at_yahoo.de)
Date: 08/09/04

  • Next message: Alexander Burger: "Re: Lisp in hardware"
    Date: Mon, 09 Aug 2004 20:48:26 +0200
    
    

    Robert Maas wrote:

    > You didn't answer my question. If you must accept all e-mail, including
    > spam, and *then* filter it locally, what do you propose the filter do
    > with all the apparent-spam it finds? Pick from one of the three choices
    > I listed above. What my CMUCL-based software currently does is bounce
    > the apparent-spam back to the spam-complaint address for whatever the
    > last-IP number was.

    You know of Paul Grahams nice paper, don't you?

       http://www.paulgraham.com/spam.html
       (or other papers from http://www.paulgraham.com/antispam.html)

    Of course that doesn't help with your question, but I think bouncing them
    back is a bad idea for the same reason it is a bad idea to reply to any
    spam mail. The spammer then get's the address verified and can sell it for
    a better price (or whatever spammers do with that).

    >> My setup is a SMTP proxy at my local computer, which does the
    >> filtering and is contacted by the eMail client and which contacts the
    >> server.
    >
    > But what exactly happens to all the messages that your filter considers
    > as likely spam and hence doesn't deliver into your inbox? (Pick from
    > the three options above, or explain any fourth option it does.)

    Ok, so in case you haven't read the above mentioned article(s) yet, here's
    a short summary: He creates two inboxes--spam and non-spam--and builds up
    statistical filtering information. This filter is "intelligent" meaning
    that it categorizes spam to *your* criteria. Actually you don't
    necessarily need to keep the spam once you told the system that this mail
    was spam and it updated the filter. But you could do in order to look for
    false-positives once in a month or year. But PG claims that he had less
    then 1% false positives. Also there're a lot of filters already
    implemented using his technique (bayesian filtering) for e.g. emacs as a
    mail client.

    HTH

    -ts

    -- 
          ,,
         \../   /  <<< The LISP Effect
        |_\\ _==__
    __ | |bb|   | _________________________________________________
    

  • Next message: Alexander Burger: "Re: Lisp in hardware"

    Relevant Pages

    • Re: Last Laugh! Spammer, age 35, Meets "Moscow Rules"
      ... that in my opinion the only good spammer was a dead spammer. ... they've never had to maintain mail servers or spam ... Some of them with their filtering stuff ...
      (comp.dcom.telecom)
    • MSF antispam info
      ... Spam and fraudulent e-mail messages are major issues for computer users ... Exchange Server, and Microsoft Exchange Hosted Filtering. ... and personalized spam protection while reducing false positives. ...
      (comp.mail.misc)
    • Re: xp sp2
      ... The amount of spam you get has nothing ... Whether you ever replied to a spammer, even if it was to "opt out". ... I'm not familiar with Mail Washer, but if this is a filtering ...
      (microsoft.public.windowsxp.general)
    • Re: Non English Spam
      ... Subject: Non English Spam ... Also this means that later filtering on the first Received field is ... repeatedly blocks yahoomail, craigslist, and ebay because spammers ... If they want to filter Asian charsets, ...
      (freebsd-questions)
    • Re: IP ranges used in North America, Hawaii, and Alaska?
      ... >> If you are trying to cut down on spam, ... > *Filtering* all that crap would cost a lot more. ... > That trick became useless when lists of dynamically-assigned ... The larger the network -- and presumably the more likely employees will ...
      (comp.os.linux.security)