Re[4]: [Declude.JunkMail] domain name a name

2005-02-11 Thread Pete McNeil
On Friday, February 11, 2005, 9:28:28 AM, Darin wrote:

DC Hi Pete,

DC Right... but the first few typically slip through before they're added to
DC your filters (like they would for anyone)...so we add them on the first
DC report to us as well.

I'll raise the feature request again --- as soon as I get my
flameproof suit on:

Declude should have a test/feature to delay a message by x hours if
the sender is not recognized. This gives all filtering mechanisms time
to adapt to new spam sources. Once the delay time has expired the
message is passed through as if it were new so that the presumably
updated BLs, filters, etc will have the ability to filter the message
(if needed).

To revive and put to rest past arguments about this:

Big reason not to do this: It is unforgivable and in all other ways a
bad idea to delay any message by any amount of time and huge amounts
of money or even lives may be lost if this happens.

To which I contend...

If this is the first time you have ever received a message from a
particular source then there is no expectation yet for the time to
delivery and email systems in general may impose end-to-end delays of
between minutes to hours depending upon many unknown factors at any
time (queues, down servers, down connectivity, graylisting (force
retry at first connect)).

Since only _new_ connections would be effected, this feature would go
almost un-noticed in the vast majority of cases. All other email
sources, where there is an expectation, would be passed at full speed
with normal filtering.

Also, IF you happen to be in a position where you really can't afford
to impose any delays on new messages then: A) You probably aren't
filtering anyway since that would be dangerous [ a conflict in policy
] and B) You _can_ turn it off ;-)

Those are my thoughts on that ( once again ).

_M

/M retreats to underground bunker  activates shields at full power.



---
[This E-mail was scanned for viruses by Declude Virus (http://www.declude.com)]

---
This E-mail came from the Declude.JunkMail mailing list.  To
unsubscribe, just send an E-mail to [EMAIL PROTECTED], and
type unsubscribe Declude.JunkMail.  The archives can be found
at http://www.mail-archive.com.


RE: Re[4]: [Declude.JunkMail] domain name a name

2005-02-11 Thread Colbeck, Andrew
Pete I agree with you.  Graylisting or greylisting would be a great add
on to Declude.

I've hoped for this in an MTA, but it doesn't look like CPHZ will go
that way, and since Ipswitch only adopts antispam measures that Declude
already has heh, it won't be coming from them.  SmarterMail may well
be more receptive to suggestions.

The current Declude architecture could do it however, as evidenced by
the nifty logic which implements the Overflow folder.  Instead of the
MTA rejecting the message with a 5xx error, Declude would tuck away the
new message Q*.SMD and D*.SMD in a folder, and re-scan the messages in
that folder that have aged sufficiently and take the usual actions.

This could also be a relatively easy 3rd party add-on, and is made a
little easier now that Declude 2.0 supports a HOLD with a specified
folder name.

For those who want to learn more about greylisting, I suggest this:

http://projects.puremagic.com/greylisting/

Andrew 8)

-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Pete McNeil
Sent: Friday, February 11, 2005 6:49 AM
To: Darin Cox
Subject: Re[4]: [Declude.JunkMail] domain name a name


On Friday, February 11, 2005, 9:28:28 AM, Darin wrote:

DC Hi Pete,

DC Right... but the first few typically slip through before they're 
DC added to your filters (like they would for anyone)...so we add them 
DC on the first report to us as well.

I'll raise the feature request again --- as soon as I get my flameproof
suit on:

Declude should have a test/feature to delay a message by x hours if the
sender is not recognized. This gives all filtering mechanisms time to
adapt to new spam sources. Once the delay time has expired the message
is passed through as if it were new so that the presumably updated BLs,
filters, etc will have the ability to filter the message (if needed).

To revive and put to rest past arguments about this:

Big reason not to do this: It is unforgivable and in all other ways a
bad idea to delay any message by any amount of time and huge amounts of
money or even lives may be lost if this happens.

To which I contend...

If this is the first time you have ever received a message from a
particular source then there is no expectation yet for the time to
delivery and email systems in general may impose end-to-end delays of
between minutes to hours depending upon many unknown factors at any time
(queues, down servers, down connectivity, graylisting (force retry at
first connect)).

Since only _new_ connections would be effected, this feature would go
almost un-noticed in the vast majority of cases. All other email
sources, where there is an expectation, would be passed at full speed
with normal filtering.

Also, IF you happen to be in a position where you really can't afford to
impose any delays on new messages then: A) You probably aren't filtering
anyway since that would be dangerous [ a conflict in policy ] and B) You
_can_ turn it off ;-)

Those are my thoughts on that ( once again ).

_M

/M retreats to underground bunker  activates shields at full power.



---
[This E-mail was scanned for viruses by Declude Virus
(http://www.declude.com)]

---
This E-mail came from the Declude.JunkMail mailing list.  To
unsubscribe, just send an E-mail to [EMAIL PROTECTED], and type
unsubscribe Declude.JunkMail.  The archives can be found at
http://www.mail-archive.com.
---
[This E-mail was scanned for viruses by Declude Virus (http://www.declude.com)]

---
This E-mail came from the Declude.JunkMail mailing list.  To
unsubscribe, just send an E-mail to [EMAIL PROTECTED], and
type unsubscribe Declude.JunkMail.  The archives can be found
at http://www.mail-archive.com.


RE: Re[4]: [Declude.JunkMail] domain name a name

2005-02-11 Thread Colbeck, Andrew
I meant to also add that I recently had many hours of planned downtime
on my MTA in my absolute lowest ham window - late Saturday evening
through early Sunday morning.  I saw very little spam increase once the
MTA was back up.

This tells me that the spammers have not yet implemented full MTAs that
retry their queued spam.  An MTA that tells them to try again later
(greylisting) would work well for me.

If greylisting that was configurable by hours was available to me, I
might turn it off during business hours for maximum safety.  I would
also want a feature to gather addresses/domains/IPs from my outbound
mail to create an autowhitelist*.

Andrew 8)

* http://eservicesforyou.com/ John Tolmachoff, do you still sell
AutoWhite?


-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Pete McNeil
Sent: Friday, February 11, 2005 6:49 AM
To: Darin Cox
Subject: Re[4]: [Declude.JunkMail] domain name a name


On Friday, February 11, 2005, 9:28:28 AM, Darin wrote:

DC Hi Pete,

DC Right... but the first few typically slip through before they're 
DC added to your filters (like they would for anyone)...so we add them 
DC on the first report to us as well.

I'll raise the feature request again --- as soon as I get my flameproof
suit on:

Declude should have a test/feature to delay a message by x hours if the
sender is not recognized. This gives all filtering mechanisms time to
adapt to new spam sources. Once the delay time has expired the message
is passed through as if it were new so that the presumably updated BLs,
filters, etc will have the ability to filter the message (if needed).

To revive and put to rest past arguments about this:

Big reason not to do this: It is unforgivable and in all other ways a
bad idea to delay any message by any amount of time and huge amounts of
money or even lives may be lost if this happens.

To which I contend...

If this is the first time you have ever received a message from a
particular source then there is no expectation yet for the time to
delivery and email systems in general may impose end-to-end delays of
between minutes to hours depending upon many unknown factors at any time
(queues, down servers, down connectivity, graylisting (force retry at
first connect)).

Since only _new_ connections would be effected, this feature would go
almost un-noticed in the vast majority of cases. All other email
sources, where there is an expectation, would be passed at full speed
with normal filtering.

Also, IF you happen to be in a position where you really can't afford to
impose any delays on new messages then: A) You probably aren't filtering
anyway since that would be dangerous [ a conflict in policy ] and B) You
_can_ turn it off ;-)

Those are my thoughts on that ( once again ).

_M

/M retreats to underground bunker  activates shields at full power.



---
[This E-mail was scanned for viruses by Declude Virus
(http://www.declude.com)]

---
This E-mail came from the Declude.JunkMail mailing list.  To
unsubscribe, just send an E-mail to [EMAIL PROTECTED], and type
unsubscribe Declude.JunkMail.  The archives can be found at
http://www.mail-archive.com.
---
[This E-mail was scanned for viruses by Declude Virus (http://www.declude.com)]

---
This E-mail came from the Declude.JunkMail mailing list.  To
unsubscribe, just send an E-mail to [EMAIL PROTECTED], and
type unsubscribe Declude.JunkMail.  The archives can be found
at http://www.mail-archive.com.


Re: Re[4]: [Declude.JunkMail] domain name a name

2005-02-11 Thread Mike K @ NetDotCom
Postfix with postgrey does exactly this.
Delays 5 minutes and maintains a db of subnet, sender  recipient combo.
Mike
- Original Message - 
From: Colbeck, Andrew [EMAIL PROTECTED]
To: Declude.JunkMail@declude.com
Sent: Friday, February 11, 2005 13:56
Subject: RE: Re[4]: [Declude.JunkMail] domain name a name

I meant to also add that I recently had many hours of planned downtime
on my MTA in my absolute lowest ham window - late Saturday evening
through early Sunday morning.  I saw very little spam increase once the
MTA was back up.
This tells me that the spammers have not yet implemented full MTAs that
retry their queued spam.  An MTA that tells them to try again later
(greylisting) would work well for me.
If greylisting that was configurable by hours was available to me, I
might turn it off during business hours for maximum safety.  I would
also want a feature to gather addresses/domains/IPs from my outbound
mail to create an autowhitelist*.
Andrew 8)
* http://eservicesforyou.com/ John Tolmachoff, do you still sell
AutoWhite?
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Pete McNeil
Sent: Friday, February 11, 2005 6:49 AM
To: Darin Cox
Subject: Re[4]: [Declude.JunkMail] domain name a name
On Friday, February 11, 2005, 9:28:28 AM, Darin wrote:
DC Hi Pete,
DC Right... but the first few typically slip through before they're
DC added to your filters (like they would for anyone)...so we add them
DC on the first report to us as well.
I'll raise the feature request again --- as soon as I get my flameproof
suit on:
Declude should have a test/feature to delay a message by x hours if the
sender is not recognized. This gives all filtering mechanisms time to
adapt to new spam sources. Once the delay time has expired the message
is passed through as if it were new so that the presumably updated BLs,
filters, etc will have the ability to filter the message (if needed).
To revive and put to rest past arguments about this:
Big reason not to do this: It is unforgivable and in all other ways a
bad idea to delay any message by any amount of time and huge amounts of
money or even lives may be lost if this happens.
To which I contend...
If this is the first time you have ever received a message from a
particular source then there is no expectation yet for the time to
delivery and email systems in general may impose end-to-end delays of
between minutes to hours depending upon many unknown factors at any time
(queues, down servers, down connectivity, graylisting (force retry at
first connect)).
Since only _new_ connections would be effected, this feature would go
almost un-noticed in the vast majority of cases. All other email
sources, where there is an expectation, would be passed at full speed
with normal filtering.
Also, IF you happen to be in a position where you really can't afford to
impose any delays on new messages then: A) You probably aren't filtering
anyway since that would be dangerous [ a conflict in policy ] and B) You
_can_ turn it off ;-)
Those are my thoughts on that ( once again ).
_M
/M retreats to underground bunker  activates shields at full power.

---
[This E-mail was scanned for viruses by Declude Virus
(http://www.declude.com)]
---
This E-mail came from the Declude.JunkMail mailing list.  To
unsubscribe, just send an E-mail to [EMAIL PROTECTED], and type
unsubscribe Declude.JunkMail.  The archives can be found at
http://www.mail-archive.com.
---
[This E-mail was scanned for viruses by Declude Virus 
(http://www.declude.com)]

---
This E-mail came from the Declude.JunkMail mailing list.  To
unsubscribe, just send an E-mail to [EMAIL PROTECTED], and
type unsubscribe Declude.JunkMail.  The archives can be found
at http://www.mail-archive.com.
---
[This E-mail was scanned for viruses by Declude Virus (http://www.declude.com)]
---
This E-mail came from the Declude.JunkMail mailing list.  To
unsubscribe, just send an E-mail to [EMAIL PROTECTED], and
type unsubscribe Declude.JunkMail.  The archives can be found
at http://www.mail-archive.com.


Re: Re[4]: [Declude.JunkMail] domain name a name

2005-02-11 Thread Darin Cox
Hmmm...some of our customers are constantly in contact with new personnel,
even new businesses, that they work with in a consulting role.  This
absolutely would not work for them, as the delays would be unacceptable.  In
their case, they'd rather see a few of the Rx spam messages get through than
have a delay on new ham.  Wouldn't work for me either for similar reasons.

Though it's certainly an idea if it can be turned off on a per-domain,
possibly even per-user basis.  A few of our customers might go for it, but
most would rather deal with the 0.5% that slip through.

Darin.


- Original Message - 
From: Pete McNeil [EMAIL PROTECTED]
To: Darin Cox Declude.JunkMail@declude.com
Sent: Friday, February 11, 2005 9:49 AM
Subject: Re[4]: [Declude.JunkMail] domain name a name


On Friday, February 11, 2005, 9:28:28 AM, Darin wrote:

DC Hi Pete,

DC Right... but the first few typically slip through before they're added
to
DC your filters (like they would for anyone)...so we add them on the first
DC report to us as well.

I'll raise the feature request again --- as soon as I get my
flameproof suit on:

Declude should have a test/feature to delay a message by x hours if
the sender is not recognized. This gives all filtering mechanisms time
to adapt to new spam sources. Once the delay time has expired the
message is passed through as if it were new so that the presumably
updated BLs, filters, etc will have the ability to filter the message
(if needed).

To revive and put to rest past arguments about this:

Big reason not to do this: It is unforgivable and in all other ways a
bad idea to delay any message by any amount of time and huge amounts
of money or even lives may be lost if this happens.

To which I contend...

If this is the first time you have ever received a message from a
particular source then there is no expectation yet for the time to
delivery and email systems in general may impose end-to-end delays of
between minutes to hours depending upon many unknown factors at any
time (queues, down servers, down connectivity, graylisting (force
retry at first connect)).

Since only _new_ connections would be effected, this feature would go
almost un-noticed in the vast majority of cases. All other email
sources, where there is an expectation, would be passed at full speed
with normal filtering.

Also, IF you happen to be in a position where you really can't afford
to impose any delays on new messages then: A) You probably aren't
filtering anyway since that would be dangerous [ a conflict in policy
] and B) You _can_ turn it off ;-)

Those are my thoughts on that ( once again ).

_M

/M retreats to underground bunker  activates shields at full power.



---
[This E-mail was scanned for viruses by Declude Virus
(http://www.declude.com)]

---
This E-mail came from the Declude.JunkMail mailing list.  To
unsubscribe, just send an E-mail to [EMAIL PROTECTED], and
type unsubscribe Declude.JunkMail.  The archives can be found
at http://www.mail-archive.com.

---
[This E-mail was scanned for viruses by Declude Virus (http://www.declude.com)]

---
This E-mail came from the Declude.JunkMail mailing list.  To
unsubscribe, just send an E-mail to [EMAIL PROTECTED], and
type unsubscribe Declude.JunkMail.  The archives can be found
at http://www.mail-archive.com.


RE: Re[4]: [Declude.JunkMail] domain name a name

2005-02-11 Thread Markus Gufler
I personaly agree completely with Pete's arguments. I've asked over a year
ago the first time for custom hold folders. The benefit of keep and check
again later is only one offered by custom hold folders. Fortunately v2 now
has custom hold folders.

I've also mentioned months ago what Matt said: requeuing and call Declude
with a VB-Script.

I personaly would have this functionality not to Delay a lot of legit mails
in order to catch also a few spam messages that slip trough. 
The idea is to bring out of the review range as much messages as possible.
Independently if reviewing is done by one operator or on client side. The
goal is to have as less messages as possible in this range. Both Operators
and clients are human and after days of seeing only spam in the review
folder, this humans will stop watching attentive for legit messages.

In order to avoid FP's I believe we all have configured our review range
pretty conservative and usualy find 0,00x % of false positives durring
review. So the delayed messages are at 99,99x% spam, and so it's absolutely
no problem if they are delayed several hours (cached DNS, Filter-Updates,
...)

If this idea it's worth would only show up a practical test with a report
what percentage of the review range would recieve a higher weight after some
hours...

Markus





 -Original Message-
 From: [EMAIL PROTECTED] 
 [mailto:[EMAIL PROTECTED] On Behalf Of Darin Cox
 Sent: Friday, February 11, 2005 8:19 PM
 To: Declude.JunkMail@declude.com
 Subject: Re: Re[4]: [Declude.JunkMail] domain name a name
 
 Hmmm...some of our customers are constantly in contact with 
 new personnel, even new businesses, that they work with in a 
 consulting role.  This absolutely would not work for them, as 
 the delays would be unacceptable.  In their case, they'd 
 rather see a few of the Rx spam messages get through than 
 have a delay on new ham.  Wouldn't work for me either for 
 similar reasons.
 
 Though it's certainly an idea if it can be turned off on a 
 per-domain, possibly even per-user basis.  A few of our 
 customers might go for it, but most would rather deal with 
 the 0.5% that slip through.
 
 Darin.
 
 
 - Original Message -
 From: Pete McNeil [EMAIL PROTECTED]
 To: Darin Cox Declude.JunkMail@declude.com
 Sent: Friday, February 11, 2005 9:49 AM
 Subject: Re[4]: [Declude.JunkMail] domain name a name
 
 
 On Friday, February 11, 2005, 9:28:28 AM, Darin wrote:
 
 DC Hi Pete,
 
 DC Right... but the first few typically slip through before 
 they're added
 to
 DC your filters (like they would for anyone)...so we add 
 them on the first
 DC report to us as well.
 
 I'll raise the feature request again --- as soon as I get my
 flameproof suit on:
 
 Declude should have a test/feature to delay a message by x hours if
 the sender is not recognized. This gives all filtering mechanisms time
 to adapt to new spam sources. Once the delay time has expired the
 message is passed through as if it were new so that the presumably
 updated BLs, filters, etc will have the ability to filter the message
 (if needed).
 
 To revive and put to rest past arguments about this:
 
 Big reason not to do this: It is unforgivable and in all other ways a
 bad idea to delay any message by any amount of time and huge amounts
 of money or even lives may be lost if this happens.
 
 To which I contend...
 
 If this is the first time you have ever received a message from a
 particular source then there is no expectation yet for the time to
 delivery and email systems in general may impose end-to-end delays of
 between minutes to hours depending upon many unknown factors at any
 time (queues, down servers, down connectivity, graylisting (force
 retry at first connect)).
 
 Since only _new_ connections would be effected, this feature would go
 almost un-noticed in the vast majority of cases. All other email
 sources, where there is an expectation, would be passed at full speed
 with normal filtering.
 
 Also, IF you happen to be in a position where you really can't afford
 to impose any delays on new messages then: A) You probably aren't
 filtering anyway since that would be dangerous [ a conflict in policy
 ] and B) You _can_ turn it off ;-)
 
 Those are my thoughts on that ( once again ).
 
 _M
 
 /M retreats to underground bunker  activates shields at full power.
 
 
 
 ---
 [This E-mail was scanned for viruses by Declude Virus
 (http://www.declude.com)]
 
 ---
 This E-mail came from the Declude.JunkMail mailing list.  To
 unsubscribe, just send an E-mail to [EMAIL PROTECTED], and
 type unsubscribe Declude.JunkMail.  The archives can be found
 at http://www.mail-archive.com.
 
 ---
 [This E-mail was scanned for viruses by Declude Virus 
 (http://www.declude.com)]
 
 ---
 This E-mail came from the Declude.JunkMail mailing list.  To
 unsubscribe, just send an E-mail to [EMAIL PROTECTED], and
 type unsubscribe Declude.JunkMail.  The archives can be found
 at http://www.mail-archive.com.
 

---
[This E-mail