TB 1.5rc2 vs TB 1.06...should I upgrade? (Mainly, re Spam filtering)

Discussion in 'Firefox' started by Jim, Jan 10, 2006.

  1. Jim

    Jim Guest

    I am running Tbird 1.06. I have read the "what's new" section
    regarding 1.5. I didn't notice anything about improvement in spam
    filtering. Has there been any? I always mark junk as junk if the
    filter does not recongize it, but recently the percentage of junk
    making it past the filter is increasing so any improved filter would
    be a plus. Is relearning required if I upgrade? And should I
    uninstall 1.06 before installing 1.5?
    Jim, Jan 10, 2006
    #1
    1. Advertising

  2. Jim

    Ed Mullen Guest

    Jim wrote:
    > I am running Tbird 1.06. I have read the "what's new" section
    > regarding 1.5. I didn't notice anything about improvement in spam
    > filtering. Has there been any? I always mark junk as junk if the
    > filter does not recongize it, but recently the percentage of junk
    > making it past the filter is increasing so any improved filter would
    > be a plus. Is relearning required if I upgrade? And should I
    > uninstall 1.06 before installing 1.5?


    Relearning is not required after an upgrade. Your junk mail database is
    stored in your profile which is not touched by a program upgrade.

    As for accuracy of the filter, spammers are constantly reformulating
    their spam to try to be more effective and outguess filters, including
    the Bayesian-type filter used by Mozilla applications. This means (at
    least) that filter training is an ongoing process.

    Every once in a while I notice the filter efficiency drop because of the
    changing spammer tactics. You can rename your training.dat file and
    train a new one from scratch. Also, I use a combination the junk mail
    filter and my ISP's spam filter which also needs to be periodically
    re-trained. When that happens I take about a week and leave messages on
    the POP server and mass-mark messages as spam. That gets two trained
    filters working on the problem. This seems, over time, to have been the
    most effective approach for me.

    --
    Ed Mullen
    http://edmullen.net
    http://mozilla.edmullen.net
    http://abington.edmullen.net
    Ed Mullen, Jan 10, 2006
    #2
    1. Advertising

  3. Jim

    Jim Guest

    On Tue, 10 Jan 2006 13:06:58 -0500, Ed Mullen <> wrote:

    >Jim wrote:
    >> I am running Tbird 1.06. I have read the "what's new" section
    >> regarding 1.5. I didn't notice anything about improvement in spam
    >> filtering. Has there been any? I always mark junk as junk if the
    >> filter does not recongize it, but recently the percentage of junk
    >> making it past the filter is increasing so any improved filter would
    >> be a plus. Is relearning required if I upgrade? And should I
    >> uninstall 1.06 before installing 1.5?

    >
    >Relearning is not required after an upgrade. Your junk mail database is
    >stored in your profile which is not touched by a program upgrade.
    >
    >As for accuracy of the filter, spammers are constantly reformulating
    >their spam to try to be more effective and outguess filters, including
    >the Bayesian-type filter used by Mozilla applications. This means (at
    >least) that filter training is an ongoing process.
    >
    >Every once in a while I notice the filter efficiency drop because of the
    >changing spammer tactics. You can rename your training.dat file and
    >train a new one from scratch. Also, I use a combination the junk mail
    >filter and my ISP's spam filter which also needs to be periodically
    >re-trained. When that happens I take about a week and leave messages on
    >the POP server and mass-mark messages as spam. That gets two trained
    >filters working on the problem. This seems, over time, to have been the
    >most effective approach for me.


    Thanks for the suggestions in the 2nd paragraph. I might try renaming
    the training.dat file.

    As for training the ISP's (Comcast) Spam Filter, I have tried that for
    a period of time when the spam began to increase. But then I was
    training two filters on the same messages: I would log on via the web,
    mark (but not delete) the spam, then download it and mark in T-bird.
    This seemed not very time-effective. Also..Comcast's (Brightmail)
    filter doesn't seem to learn anything by my marking spam, their "thank
    you" message not withstanding.
    Jim, Jan 10, 2006
    #3
  4. Jim

    DoubleDoom Guest

    Jim wrote:
    > I am running Tbird 1.06. I have read the "what's new" section
    > regarding 1.5. I didn't notice anything about improvement in spam
    > filtering. Has there been any? I always mark junk as junk if the
    > filter does not recongize it, but recently the percentage of junk
    > making it past the filter is increasing so any improved filter would
    > be a plus. Is relearning required if I upgrade? And should I
    > uninstall 1.06 before installing 1.5?


    Why don't you try a custom build of TB? It doesn't overwrite your 1.06
    and you can test the new version out. If you don't like it, you delete
    the folder and revert to the old one.

    You can get a custom build from here:

    http://www.vector64.com/WindowsBuilds.html

    If you want to be ultra safe, you can always backup your profile first.

    David
    DoubleDoom, Jan 10, 2006
    #4
  5. Jim

    Ed Mullen Guest

    Jim wrote:
    > On Tue, 10 Jan 2006 13:06:58 -0500, Ed Mullen <> wrote:
    >
    >> Jim wrote:
    >>> I am running Tbird 1.06. I have read the "what's new" section
    >>> regarding 1.5. I didn't notice anything about improvement in spam
    >>> filtering. Has there been any? I always mark junk as junk if the
    >>> filter does not recongize it, but recently the percentage of junk
    >>> making it past the filter is increasing so any improved filter would
    >>> be a plus. Is relearning required if I upgrade? And should I
    >>> uninstall 1.06 before installing 1.5?

    >> Relearning is not required after an upgrade. Your junk mail database is
    >> stored in your profile which is not touched by a program upgrade.
    >>
    >> As for accuracy of the filter, spammers are constantly reformulating
    >> their spam to try to be more effective and outguess filters, including
    >> the Bayesian-type filter used by Mozilla applications. This means (at
    >> least) that filter training is an ongoing process.
    >>
    >> Every once in a while I notice the filter efficiency drop because of the
    >> changing spammer tactics. You can rename your training.dat file and
    >> train a new one from scratch. Also, I use a combination the junk mail
    >> filter and my ISP's spam filter which also needs to be periodically
    >> re-trained. When that happens I take about a week and leave messages on
    >> the POP server and mass-mark messages as spam. That gets two trained
    >> filters working on the problem. This seems, over time, to have been the
    >> most effective approach for me.

    >
    > Thanks for the suggestions in the 2nd paragraph. I might try renaming
    > the training.dat file.
    >
    > As for training the ISP's (Comcast) Spam Filter, I have tried that for
    > a period of time when the spam began to increase. But then I was
    > training two filters on the same messages: I would log on via the web,
    > mark (but not delete) the spam, then download it and mark in T-bird.
    > This seemed not very time-effective. Also..Comcast's (Brightmail)
    > filter doesn't seem to learn anything by my marking spam, their "thank
    > you" message not withstanding.


    First, my approach is to use SeaMonkey (in your case TB) to retrieve,
    read, reply to email. When I'm confronted with the need to "train" the
    Comcast filter I temporarily (for a couple weeks?) set SeaMonkey to
    leave all message on the POP server (Comcast) for 7 days, after which
    the are automatically deleted in case I get busy and don't' have time
    (or forget!) to go to the Webmail interface and deal with it. This
    prevents my server storage from being taxed and gives me a week to
    occasionally log onto the Comcast Webmail and mark things as Spam.
    Also, I don't have to worry about not deleting "good" mail on the
    Webmail interface; it's already on my local machine and has been dealt
    with appropriately (saved, replied to, deleted, etc.). It's just a
    matter of cleaning up the Comcast server so it doesn't get overloaded.
    When I check various messages there and click the "Spam" button the
    messages are automatically deleted.

    Yes, it's a bit more work than just relying on the SM/TB junk filter but
    I'm not doing it all day long, every day, just once a week, or maybe
    every couple of days. Same effect that your method has but less trouble.

    As to the effectiveness of Comcast's filter, all I can say is that
    casual observation (vs. scientific testing/analysis) indicates to me
    that it does work. The problem is in the adaptive abilities of the
    spammers. That's why this is a never-ending battle. ;-)

    --
    Ed Mullen
    http://edmullen.net
    http://mozilla.edmullen.net
    http://abington.edmullen.net
    Ed Mullen, Jan 10, 2006
    #5
  6. Jim

    Jim Guest

    On Tue, 10 Jan 2006 16:31:06 -0500, Ed Mullen <> wrote:

    >Jim wrote:
    >> On Tue, 10 Jan 2006 13:06:58 -0500, Ed Mullen <> wrote:
    >>
    >>> Jim wrote:
    >>>> I am running Tbird 1.06. I have read the "what's new" section
    >>>> regarding 1.5. I didn't notice anything about improvement in spam
    >>>> filtering. Has there been any? I always mark junk as junk if the
    >>>> filter does not recongize it, but recently the percentage of junk
    >>>> making it past the filter is increasing so any improved filter would
    >>>> be a plus. Is relearning required if I upgrade? And should I
    >>>> uninstall 1.06 before installing 1.5?
    >>> Relearning is not required after an upgrade. Your junk mail database is
    >>> stored in your profile which is not touched by a program upgrade.
    >>>
    >>> As for accuracy of the filter, spammers are constantly reformulating
    >>> their spam to try to be more effective and outguess filters, including
    >>> the Bayesian-type filter used by Mozilla applications. This means (at
    >>> least) that filter training is an ongoing process.
    >>>
    >>> Every once in a while I notice the filter efficiency drop because of the
    >>> changing spammer tactics. You can rename your training.dat file and
    >>> train a new one from scratch. Also, I use a combination the junk mail
    >>> filter and my ISP's spam filter which also needs to be periodically
    >>> re-trained. When that happens I take about a week and leave messages on
    >>> the POP server and mass-mark messages as spam. That gets two trained
    >>> filters working on the problem. This seems, over time, to have been the
    >>> most effective approach for me.

    >>
    >> Thanks for the suggestions in the 2nd paragraph. I might try renaming
    >> the training.dat file.
    >>
    >> As for training the ISP's (Comcast) Spam Filter, I have tried that for
    >> a period of time when the spam began to increase. But then I was
    >> training two filters on the same messages: I would log on via the web,
    >> mark (but not delete) the spam, then download it and mark in T-bird.
    >> This seemed not very time-effective. Also..Comcast's (Brightmail)
    >> filter doesn't seem to learn anything by my marking spam, their "thank
    >> you" message not withstanding.

    >
    >First, my approach is to use SeaMonkey (in your case TB) to retrieve,
    >read, reply to email. When I'm confronted with the need to "train" the
    >Comcast filter I temporarily (for a couple weeks?) set SeaMonkey to
    >leave all message on the POP server (Comcast) for 7 days, after which
    >the are automatically deleted in case I get busy and don't' have time
    >(or forget!) to go to the Webmail interface and deal with it. This
    >prevents my server storage from being taxed and gives me a week to
    >occasionally log onto the Comcast Webmail and mark things as Spam.
    >Also, I don't have to worry about not deleting "good" mail on the
    >Webmail interface; it's already on my local machine and has been dealt
    >with appropriately (saved, replied to, deleted, etc.). It's just a
    >matter of cleaning up the Comcast server so it doesn't get overloaded.
    >When I check various messages there and click the "Spam" button the
    >messages are automatically deleted.
    >
    >Yes, it's a bit more work than just relying on the SM/TB junk filter but
    >I'm not doing it all day long, every day, just once a week, or maybe
    >every couple of days. Same effect that your method has but less trouble.
    >
    >As to the effectiveness of Comcast's filter, all I can say is that
    >casual observation (vs. scientific testing/analysis) indicates to me
    >that it does work. The problem is in the adaptive abilities of the
    >spammers. That's why this is a never-ending battle. ;-)


    I suppose Comcast's filter does work pretty well, since it averages 50
    spams a day caught vs a dozen or so missed. Fortunately, only problem
    is in my primary account.

    I do like your suggestion re: leaving stuff on the Comcast Server and
    just logging on once in a while. No need to worry about exceeding 250
    mb.
    Jim, Jan 11, 2006
    #6
    1. Advertising

Want to reply to this thread or ask your own question?

It takes just 2 minutes to sign up (and it's free!). Just click the sign up button to choose a username and then you can ask your own questions on the forum.
Similar Threads
  1. Replies:
    4
    Views:
    953
    dantu
    Oct 14, 2003
  2. Brian H¹©
    Replies:
    2
    Views:
    563
    Brian H¹©
    Jul 18, 2003
  3. Jersey
    Replies:
    2
    Views:
    411
    Jersey
    Jul 23, 2004
  4. The Great Cornholio
    Replies:
    0
    Views:
    383
    The Great Cornholio
    Jul 23, 2004
  5. Josef Griffiths

    Searching Mainly Brings Up "Other Search Engines"

    Josef Griffiths, Aug 2, 2004, in forum: Computer Support
    Replies:
    9
    Views:
    632
    °Mike°
    Aug 3, 2004
Loading...

Share This Page