Speaking of cgf's enhancement to the rbl checker...

Speaking of cgf's enhancement to the rbl checker...

[Jason Molenda]

I was curious what percentage of people were avoiding the RBL checks
by being subscribed.

Since January 1st, rbl-check.sh has consulted the actual RBLs 10,127
times.  It has skipped these RBL checks 5,883 times because people
were subscribed.  Of the 10,127 RBL checks, 9,672 times the messages
were let through.  (the remainder were rejected because the sending
site was listed in the RBLs)

Of course, this doesn't tell the whole story - a small number of
people make up much of those 10,127 RBL checks.  The top 20:

   1204 anonymous@sourceware.cygnus.com
    472 aoliva@redhat.com
    324 tromey@creche.cygnus.com
    265 mitchell@codesourcery.com
    166 bryce@albatross.co.nz
    163 pedwards@disaster.jaj.com
    159 law@upchuck.cygnus.com
    144 jsm28@srcf.ucam.org
    139 drepper@redhat.com
    135 ghazi@caip.rutgers.edu
    130 jlarmour@redhat.com
    127 dave@hiauly1.hia.nrc.ca
    120 nickc@redhat.com
    105 jason@redhat.com
    103 kenner@vlsi1.ultra.nyu.edu
    103 cwilson@ece.gatech.edu
     93 corinna@vinschen.de
     89 cgf@redhat.com
     86 hubicka@atrey.karlin.mff.cuni.cz
     74 bernds@redhat.com

Or about 4,201 of those checks in all.  So if you exclude these
folks, the RBL checks have been made unnecessary for about 50% of
all senders.

Jason

Re: Speaking of cgf's enhancement to the rbl checker...

[Jason Molenda]

Oh BTW the breakdown on which RBLs were giving the positive
identifications.  since Jan 1, they are

    273 RSS
     41 MAPS DUL
     25 ORBS
      5 MAPS RBL

It might be interesting to flag in rbl-check.sh which of the ORBS
lists a site is listed in.

J

Re: Speaking of cgf's enhancement to the rbl checker...

[Gerald Pfeifer]

On Wed, 7 Feb 2001, Jason Molenda wrote:
> Oh BTW the breakdown on which RBLs were giving the positive
> identifications.  since Jan 1, they are

I assume this list is not ``fair'' in that if we have a block from
the first RBL, we won't consult any others?

>     273 RSS
>      41 MAPS DUL
>      25 ORBS
>       5 MAPS RBL

Intuitively (and according to my experience), if you switch RSS and ORBS,
you'd get a symmetric result, right?

Gerald
-- 
Gerald "Jerry" pfeifer@dbai.tuwien.ac.at http://www.dbai.tuwien.ac.at/~pfeifer/

Re: Speaking of cgf's enhancement to the rbl checker...

[Jason Molenda]

On Thu, Feb 08, 2001 at 02:47:04AM +0100, Gerald Pfeifer wrote:
> On Wed, 7 Feb 2001, Jason Molenda wrote:
> > Oh BTW the breakdown on which RBLs were giving the positive
> > identifications.  since Jan 1, they are
> 
> I assume this list is not ``fair'' in that if we have a block from
> the first RBL, we won't consult any others?

Good point - that is the case.  I think the order (this is from long-ago
memories) is MAPS, MAPS DUL, RSS, ORBS.

I was wondering why the ORBS number was so small. :-)

> Intuitively (and according to my experience), if you switch RSS and ORBS,
> you'd get a symmetric result, right?

I haven't looked in a long time, but it used to be like that.  I
don't know if it's still true.


J

Re: Speaking of cgf's enhancement to the rbl checker...

[Christopher Faylor]

On Wed, Feb 07, 2001 at 05:55:31PM -0800, Jason Molenda wrote:
>> Intuitively (and according to my experience), if you switch RSS and ORBS,
>> you'd get a symmetric result, right?
>
>I haven't looked in a long time, but it used to be like that.  I
>don't know if it's still true.

When I was playing with things last weekend, I found some situations where
ORBS blocked things that RSS didn't but not vice versa.

cgf

Re: Speaking of cgf's enhancement to the rbl checker...

[Alexandre Oliva]

On Feb  7, 2001, Jason Molenda <jason-swarelist@molenda.com> wrote:

> I was curious what percentage of people were avoiding the RBL checks
> by being subscribed.

>    1204 anonymous@sourceware.cygnus.com
>     472 aoliva@redhat.com

Sorry about all the SPAM :-)

Now, on to figure out who this anonymous@ user is, and terminate
him/her, so that I can claim the prize :-)

-- 
Alexandre Oliva   Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer                  aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicamp        oliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist    *Please* write to mailing lists, not to me

Re: Speaking of cgf's enhancement to the rbl checker...

[Jason Molenda]

On Thu, Feb 08, 2001 at 07:11:45PM -0200, Alexandre Oliva wrote:

> >    1204 anonymous@sourceware.cygnus.com
> >     472 aoliva@redhat.com
> 
> Sorry about all the SPAM :-)
> 
> Now, on to figure out who this anonymous@ user is, and terminate
> him/her, so that I can claim the prize :-)

:)  This is probably the gnats mail notes or something like that.
cvs messages go through a separate script (local-mail-only.sh).

J

Re: Speaking of cgf's enhancement to the rbl checker...

[Jason Molenda]

On Thu, Feb 08, 2001 at 02:47:04AM +0100, Gerald Pfeifer wrote:
> On Wed, 7 Feb 2001, Jason Molenda wrote:
> > Oh BTW the breakdown on which RBLs were giving the positive
> > identifications.  since Jan 1, they are
> 
> I assume this list is not ``fair'' in that if we have a block from
> the first RBL, we won't consult any others?

Good point - that is the case.  I think the order (this is from long-ago
memories) is MAPS, MAPS DUL, RSS, ORBS.

I was wondering why the ORBS number was so small. :-)

> Intuitively (and according to my experience), if you switch RSS and ORBS,
> you'd get a symmetric result, right?

I haven't looked in a long time, but it used to be like that.  I
don't know if it's still true.


J

Re: Speaking of cgf's enhancement to the rbl checker...

[Gerald Pfeifer]

On Wed, 7 Feb 2001, Jason Molenda wrote:
> Oh BTW the breakdown on which RBLs were giving the positive
> identifications.  since Jan 1, they are

I assume this list is not ``fair'' in that if we have a block from
the first RBL, we won't consult any others?

>     273 RSS
>      41 MAPS DUL
>      25 ORBS
>       5 MAPS RBL

Intuitively (and according to my experience), if you switch RSS and ORBS,
you'd get a symmetric result, right?

Gerald
-- 
Gerald "Jerry" pfeifer@dbai.tuwien.ac.at http://www.dbai.tuwien.ac.at/~pfeifer/

Speaking of cgf's enhancement to the rbl checker...

[Jason Molenda]

I was curious what percentage of people were avoiding the RBL checks
by being subscribed.

Since January 1st, rbl-check.sh has consulted the actual RBLs 10,127
times.  It has skipped these RBL checks 5,883 times because people
were subscribed.  Of the 10,127 RBL checks, 9,672 times the messages
were let through.  (the remainder were rejected because the sending
site was listed in the RBLs)

Of course, this doesn't tell the whole story - a small number of
people make up much of those 10,127 RBL checks.  The top 20:

   1204 anonymous@sourceware.cygnus.com
    472 aoliva@redhat.com
    324 tromey@creche.cygnus.com
    265 mitchell@codesourcery.com
    166 bryce@albatross.co.nz
    163 pedwards@disaster.jaj.com
    159 law@upchuck.cygnus.com
    144 jsm28@srcf.ucam.org
    139 drepper@redhat.com
    135 ghazi@caip.rutgers.edu
    130 jlarmour@redhat.com
    127 dave@hiauly1.hia.nrc.ca
    120 nickc@redhat.com
    105 jason@redhat.com
    103 kenner@vlsi1.ultra.nyu.edu
    103 cwilson@ece.gatech.edu
     93 corinna@vinschen.de
     89 cgf@redhat.com
     86 hubicka@atrey.karlin.mff.cuni.cz
     74 bernds@redhat.com

Or about 4,201 of those checks in all.  So if you exclude these
folks, the RBL checks have been made unnecessary for about 50% of
all senders.

Jason

Re: Speaking of cgf's enhancement to the rbl checker...

[Jason Molenda]

Oh BTW the breakdown on which RBLs were giving the positive
identifications.  since Jan 1, they are

    273 RSS
     41 MAPS DUL
     25 ORBS
      5 MAPS RBL

It might be interesting to flag in rbl-check.sh which of the ORBS
lists a site is listed in.

J

Re: Speaking of cgf's enhancement to the rbl checker...

[Alexandre Oliva]

On Feb  7, 2001, Jason Molenda <jason-swarelist@molenda.com> wrote:

> I was curious what percentage of people were avoiding the RBL checks
> by being subscribed.

>    1204 anonymous@sourceware.cygnus.com
>     472 aoliva@redhat.com

Sorry about all the SPAM :-)

Now, on to figure out who this anonymous@ user is, and terminate
him/her, so that I can claim the prize :-)

-- 
Alexandre Oliva   Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer                  aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicamp        oliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist    *Please* write to mailing lists, not to me

Re: Speaking of cgf's enhancement to the rbl checker...

[Jason Molenda]

On Thu, Feb 08, 2001 at 07:11:45PM -0200, Alexandre Oliva wrote:

> >    1204 anonymous@sourceware.cygnus.com
> >     472 aoliva@redhat.com
> 
> Sorry about all the SPAM :-)
> 
> Now, on to figure out who this anonymous@ user is, and terminate
> him/her, so that I can claim the prize :-)

:)  This is probably the gnats mail notes or something like that.
cvs messages go through a separate script (local-mail-only.sh).

J

Re: Speaking of cgf's enhancement to the rbl checker...

[Christopher Faylor]

On Wed, Feb 07, 2001 at 05:55:31PM -0800, Jason Molenda wrote:
>> Intuitively (and according to my experience), if you switch RSS and ORBS,
>> you'd get a symmetric result, right?
>
>I haven't looked in a long time, but it used to be like that.  I
>don't know if it's still true.

When I was playing with things last weekend, I found some situations where
ORBS blocked things that RSS didn't but not vice versa.

cgf