Discussion:
Corrupt Header
Vince Puzzella
2002-07-11 15:41:03 UTC
Permalink
I'm in the process of integrating spamassassin with zmailer.

After running:
spamassassin -t < origmessage.txt > screwedupmessage.txt

The screwedupmessage.txt file contains an additional line at the top
This is corrupting our messages. Does anyone else have this problem or
any ideas on how to prevent this?

Thanks in advance.

v i n c e p u z z e l l a
s o f t w a r e d e v e l o p e r
(905) 762-5229
http://bluecatnetworks.com
Theo Van Dinter
2002-07-11 15:58:28 UTC
Permalink
Post by Vince Puzzella
The screwedupmessage.txt file contains an additional line at the top
This is corrupting our messages. Does anyone else have this problem or
any ideas on how to prevent this?
It's an mbox message seperator line. You can turn off the generation
when running SA by using the "-F 0" flag.
--
Randomly Generated Tagline:
I'd love to, but I'm having all my plants neutered.


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
PC Mods, Computing goodies, cases & more
http://thinkgeek.com/sf
Vince Puzzella
2002-07-11 17:08:40 UTC
Permalink
Thanks, that worked. BTW, what is it's purpose and why is it generated
by default? Just curious.

v i n c e p u z z e l l a
s o f t w a r e d e v e l o p e r
(905) 762-5229
http://bluecatnetworks.com


-----Original Message-----
From: Theo Van Dinter [mailto:***@kluge.net]
Sent: Thursday, July 11, 2002 11:58 AM
To: Vince Puzzella
Cc: spamassassin-***@lists.sourceforge.net
Subject: Re: [SAtalk] Corrupt Header
Post by Vince Puzzella
The screwedupmessage.txt file contains an additional line at the top
This is corrupting our messages. Does anyone else have this problem or
any ideas on how to prevent this?
It's an mbox message seperator line. You can turn off the generation
when running SA by using the "-F 0" flag.
--
Randomly Generated Tagline:
I'd love to, but I'm having all my plants neutered.


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
PC Mods, Computing goodies, cases & more
http://thinkgeek.com/sf
Theo Van Dinter
2002-07-11 17:17:29 UTC
Permalink
Post by Vince Puzzella
Thanks, that worked. BTW, what is it's purpose and why is it generated
by default? Just curious.
Most people use mbox style mail folders, the "From " at the start of a
line indicates a new message in the folder, so (indented):

From ***@foo ...
Message headers ...

Message body

From ***@foo ...
Message headers ...

Message body

Since most people use them, the default is to add the separator.

Although in thinking about it, people don't really use SA for delivery
any more, so we should probably just not add the From line and let
procmail (or insert your favorite program here) handle the From line.
Not my call though.
--
Randomly Generated Tagline:
"Why does Sea World have a seafood restaurant? I'm halfway through my
fishburger and I realize, Oh my God....I could be eating a slow
learner." - Lynda Montgomery


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
PC Mods, Computing goodies, cases & more
http://thinkgeek.com/sf
Malte S. Stretz
2002-07-12 00:14:24 UTC
Permalink
Post by Theo Van Dinter
Post by Vince Puzzella
Thanks, that worked. BTW, what is it's purpose and why is it generated
by default? Just curious.
[...]
Although in thinking about it, people don't really use SA for delivery
any more, so we should probably just not add the From line and let
procmail (or insert your favorite program here) handle the From line.
[...]
Full ACK :o)

Malte
--
-- Coding is art.
--
-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
PC Mods, Computing goodies, cases & more
http://thinkgeek.com/sf
Craig R.Hughes
2002-07-12 16:12:17 UTC
Permalink
I've long been of the philosophy that spamassassin should focus
on exactly one thing: identification and tagging of spam. Leave
the delivery to something which does delivery well. Leave the
SMTP to something which does SMTP well. etc, etc, etc. I
haven't chopped out the delivery stuff altogether, but have been
sorely tempted to.

C
Post by Malte S. Stretz
Post by Theo Van Dinter
Post by Vince Puzzella
Thanks, that worked. BTW, what is it's purpose and why is it generated
by default? Just curious.
[...]
Although in thinking about it, people don't really use SA for delivery
any more, so we should probably just not add the From line and let
procmail (or insert your favorite program here) handle the From line.
[...]
Full ACK :o)
Malte
--
-- Coding is art.
--
-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
PC Mods, Computing goodies, cases & more
http://thinkgeek.com/sf
_______________________________________________
Spamassassin-talk mailing list
https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Gadgets, caffeine, t-shirts, fun stuff.
http://thinkgeek.com/sf
Derrick 'dman' Hudson
2002-07-13 00:07:31 UTC
Permalink
On Fri, Jul 12, 2002 at 09:12:17AM -0700, Craig R. Hughes wrote:
| I've long been of the philosophy that spamassassin should focus
| on exactly one thing: identification and tagging of spam.

YES!

| Leave the delivery to something which does delivery well.

YES!

| Leave the SMTP to something which does SMTP well.

YES!

| etc, etc, etc.

"Do one thing and do it well" -- the UNIX Way.

| I haven't chopped out the delivery stuff altogether, but have been
| sorely tempted to.

Do it! Do it!

You don't know how many times I've had to help people through mail
loss/corruption due to the default From_ line adding hosing their
delivery system and mbox, and then explain what SA was trying to do
and how to stop it from misbehaving.

I will gladly welcome the removal of that misfeature.

-D
--
I tell you the truth, everyone who sins is a slave to sin. Now a slave
has no permanent place in the family, but a son belongs to it forever.
So if the Son sets you free, you will be free indeed.
John 8:34-36

http://dman.ddts.net/~dman/
Theo Van Dinter
2002-07-13 00:13:46 UTC
Permalink
Post by Derrick 'dman' Hudson
| I haven't chopped out the delivery stuff altogether, but have been
| sorely tempted to.
Do it! Do it!
I'm hoping to get a patch ready this evening. :) Although there still
needs to be a bug put in about it.
Post by Derrick 'dman' Hudson
You don't know how many times I've had to help people through mail
loss/corruption due to the default From_ line adding hosing their
delivery system and mbox, and then explain what SA was trying to do
and how to stop it from misbehaving.
Unfortunately, I think we'll still have this problem. procmail doesn't
care if SA tries to add the From header, even if it already exists.
So we'll probably still have the rom issue.
--
Randomly Generated Tagline:
"A: You mean it's not due on Friday?
T: No, it's due two Mondays from now ... <slight pause>
A: Where's the remote?!?!?" - Alan and Theo about an OS1 Assignment
Theo Van Dinter
2002-07-12 16:37:59 UTC
Permalink
Post by Theo Van Dinter
Although in thinking about it, people don't really use SA for delivery
any more, so we should probably just not add the From line and let
procmail (or insert your favorite program here) handle the From line.
Not my call though.
Ugh. I just noticed that procmail apparently doesn't add the "From "
line, which confuses the heck out of me since sendmail ought not to do
it, that's the delivery agent's job.

I found this out by changing the spamd options to have a "-F 0".
spamc gets called during procmail which promptly puts messages w/out
"From " lines in the last message of my inbox.

I'll have to dig around to figure out why it doesn't do what I expect.
The man page says it should:

When in explicit delivery mode, procmail will generate a leading `From '
line if none is present. If one is already present procmail will leave
it intact.
--
Randomly Generated Tagline:
*}- Tribble Olympics: Archery


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Gadgets, caffeine, t-shirts, fun stuff.
http://thinkgeek.com/sf
Ryan Cleary
2002-07-12 17:57:42 UTC
Permalink
Post by Theo Van Dinter
I found this out by changing the spamd options to have a "-F 0".
spamc gets called during procmail which promptly puts messages w/out
"From " lines in the last message of my inbox.
I think spamassassin is stripping the From lines. From spamassassin's
manpage:

-F 0 | 1, --add-from, --no-add-from
Ensure that the output email message either always
starts with a 'From ' line (1) for UNIX mbox format,
or ensure that this line is stripped from the output
(0). (default: add)

"ensure that this line is stripped"

-Ryan
--
T Ryan Cleary <***@interdimensions.com>
URL: http://people.interdimensions.com/tryanc
PGP: 82 93 32 D7 3A AC C0 8D 34 56 96 CC DA DB 5E 2B



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Gadgets, caffeine, t-shirts, fun stuff.
http://thinkgeek.com/sf
Theo Van Dinter
2002-07-12 21:20:42 UTC
Permalink
Post by Ryan Cleary
"ensure that this line is stripped"
Ok, I think it's time to send in a patch that gets rid of this stuff
once and for all. Let your delivery agent figure out about dealing with
the From.
--
Randomly Generated Tagline:
"A committee is a life form with 6 or more legs and no brain."
- Robert Heinlein


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Gadgets, caffeine, t-shirts, fun stuff.
http://thinkgeek.com/sf
Smart, Dan
2002-07-11 17:58:51 UTC
Permalink
I do this in Procmail with the following right before copying tagged
messages to MBOX files for logging.

----- snip -- procmailrc ---------

# Put From at beginning for MBOX format so Pine can read
:0 fhw
| formail -I "From " -a "From "

----- snip -------

<<Dan>>

|-----Original Message-----
|From: Theo Van Dinter [mailto:***@kluge.net]
|Sent: Thursday, July 11, 2002 12:17 PM
|To: Vince Puzzella
|Cc: spamassassin-***@lists.sourceforge.net
|Subject: Re: [SAtalk] Corrupt Header
|
|
|On Thu, Jul 11, 2002 at 01:08:40PM -0400, Vince Puzzella wrote:
|> Thanks, that worked. BTW, what is it's purpose and why is
|it generated
|> by default? Just curious.
|
|Most people use mbox style mail folders, the "From " at the start of a
|line indicates a new message in the folder, so (indented):
|
| From ***@foo ...
| Message headers ...
|
| Message body
|
| From ***@foo ...
| Message headers ...
|
| Message body
|
|Since most people use them, the default is to add the separator.
|
|Although in thinking about it, people don't really use SA for delivery
|any more, so we should probably just not add the From line and let
|procmail (or insert your favorite program here) handle the From line.
|Not my call though.
|
|--
|Randomly Generated Tagline:
|"Why does Sea World have a seafood restaurant? I'm halfway through my
| fishburger and I realize, Oh my God....I could be eating a slow
| learner." - Lynda Montgomery
|
|
|-------------------------------------------------------
|This sf.net email is sponsored by:ThinkGeek
|PC Mods, Computing goodies, cases & more
|http://thinkgeek.com/sf
|_______________________________________________
|Spamassassin-talk mailing list
|Spamassassin-***@lists.sourceforge.net
|https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
|


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
PC Mods, Computing goodies, cases & more
http://thinkgeek.com/sf
Justin Mason
2002-07-12 16:21:49 UTC
Permalink
Post by Craig R.Hughes
I've long been of the philosophy that spamassassin should focus
on exactly one thing: identification and tagging of spam. Leave
the delivery to something which does delivery well. Leave the
SMTP to something which does SMTP well. etc, etc, etc. I
haven't chopped out the delivery stuff altogether, but have been
sorely tempted to.
Well, at this stage I think it's fair enough to retire delivery support.
Everyone who's doing this should be using procmail, or Mail::Audit.

I suggest for the next release we flag this right at the top of the
README, download page, and release notices, and in the code just spit out
a warning and "exit 75" (to avoid lost mail).

Anyone object?

--j.


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Gadgets, caffeine, t-shirts, fun stuff.
http://thinkgeek.com/sf
Justin Mason
2002-07-30 12:02:20 UTC
Permalink
titties, tits, titlicious, etc.
I was pretty careful when I went through and created the list to
think of which ones needed to be suffix-loose and which ones
wanted to be suffix-tight. I would strongly recommend leaving
them as I had it.
But: title, titration, titan, tithes, titter... IMO it's too
easy to FP on a 3-letter word, and the FP rate was quite high
in the results I posted.

--j.


-------------------------------------------------------
This sf.net email is sponsored by: Dice - The leading online job board
for high-tech professionals. Search and apply for tech jobs today!
http://seeker.dice.com/seeker.epl?rel_code=31
Tony L. Svanstrom
2002-07-30 12:40:06 UTC
Permalink
Post by Justin Mason
titties, tits, titlicious, etc.
I was pretty careful when I went through and created the list to
think of which ones needed to be suffix-loose and which ones
wanted to be suffix-tight. I would strongly recommend leaving
them as I had it.
But: title, titration, titan, tithes, titter... IMO it's too
easy to FP on a 3-letter word, and the FP rate was quite high
in the results I posted.
[***@localhost] /Users/tony/ {154} word tit | grep ^tit | wc
web2
140 140 1376


/Tony
--
# Per scientiam ad libertatem! // Through knowledge towards freedom! #
# Genom kunskap mot frihet! =*= (c) 1999-2002 ***@svanstrom.com =*= #

perl -e'print$_{$_} for sort%_=`lynx -source svanstrom.com/t`'



-------------------------------------------------------
This sf.net email is sponsored by: Dice - The leading online job board
for high-tech professionals. Search and apply for tech jobs today!
http://seeker.dice.com/seeker.epl?rel_code=31
Lon Jones
2002-08-02 05:27:53 UTC
Permalink
Craig,
At this point I trust you (spam-assassin) to filter my e-mail.
I also trust you to do the right thing with regard to trademark,
IP and such. Please keep it up and best of luck to you and all
contributers.

Lon


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
John Goggan
2002-08-02 17:44:39 UTC
Permalink
I've had quite a few questions about the TRADEMARK file (and the
section about name use rights in the PAL). The summary is that
Deersoft owns the SpamAssassin trademark.
Did you use a third party to register your trademark? Or is "The Incentive
Company" someone else? The US Trademark Office has a relatively new filing
(May 31, 2002) for the Word Mark "Spam Assassin" registered to "The Incentive
Company CORPORATION DELAWARE 844 5th Street Lake Oswego OREGON 97034"

Note that it is still just a filed application -- it hasn't been assigned to
an examining attorney yet (as of Jun 5, 2002).

In any case, if that is you, great. If that isn't you -- then you might want
to look into it -- stopping such things before they are completed is much
easier than after... You can do lookups at http://tess.uspto.gov

- John...


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
Craig R.Hughes
2002-08-02 18:29:08 UTC
Permalink
"The Incentive Company" is a front for John Gotts. He
registered spamassassin.com shortly (a few days) after Justin's
initial SpamAssassin release announcement, and then registered
spamassassinpro.com and filed the trademark application when he
learned about Deersoft. Our lawyer is in touch with him to
rectify this situation. Needless to say, he doesn't actually
market any product called SpamAssassin, so his attempt to
register is completely bogus, and Deersoft (through Justin and
myself) have clear common law rights to the mark which pre-date
Gotts' registration. He's basically a typical nasty domain
squatter who's trying to pump up the value of the domains he's
squatting on. It's not our intention to back down and/or pay
him off though, so if he insists on making us pay our attorneys
to proceed with getting him to stop his abuses, he'll have to
pay his own attorneys a fair amount too :)

Extortion works best when there's little downside for the extorter.

We have extremely competent, very well known, trademark
attorneys working on this for us, and we've basically been all
over it for a while now. Thanks for the heads up though.

C
Post by John Goggan
I've had quite a few questions about the TRADEMARK file (and the
section about name use rights in the PAL). The summary is that
Deersoft owns the SpamAssassin trademark.
Did you use a third party to register your trademark? Or is
"The Incentive
Company" someone else? The US Trademark Office has a
relatively new filing
(May 31, 2002) for the Word Mark "Spam Assassin" registered to
"The Incentive
Company CORPORATION DELAWARE 844 5th Street Lake Oswego OREGON 97034"
Note that it is still just a filed application -- it hasn't
been assigned to
an examining attorney yet (as of Jun 5, 2002).
In any case, if that is you, great. If that isn't you -- then
you might want
to look into it -- stopping such things before they are
completed is much
easier than after... You can do lookups at http://tess.uspto.gov
-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
Christopher Crowley
2002-08-02 19:19:10 UTC
Permalink
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Here is questionable advertising that I noticed:
http://www.google.com/search?cat=&meta=&q=spamassassin


A google search on spamassassin returns:
- ----------====
SPAMASSASSIN - Stop 98% of Spam with ActiveState
www.ActiveState.com/stopspam Free evaluation. Sponsored Link
SpamAssassin - Free Trial
www.deersoft.com Stop 95% of spam with SpamAssassin Outlook
2000 & 2002/XP Sponsored Link

Category: Computers > Internet > Abuse > Spam > Filtering

...
- ----------====
Post by John Goggan
I've had quite a few questions about the TRADEMARK file (and the
section about name use rights in the PAL). The summary is that
Deersoft owns the SpamAssassin trademark.
Did you use a third party to register your trademark? Or is
"The Incentive
Company" someone else? The US Trademark Office has a
relatively new filing
(May 31, 2002) for the Word Mark "Spam Assassin" registered to
"The Incentive
Company CORPORATION DELAWARE 844 5th Street Lake Oswego OREGON 97034"
Note that it is still just a filed application -- it hasn't
been assigned to
an examining attorney yet (as of Jun 5, 2002).
-----BEGIN PGP SIGNATURE-----
Version: PGPfreeware 6.5.8 for non-commercial use <http://www.pgp.com>

iQA/AwUBPUrbLc0q7cR88Y/bEQI4owCg3ivqjRCLrZK9CRvBJajFTKQFfZkAoJly
hxDt5DN9/PTTV5T2a8mVqGpf
=tW8g
-----END PGP SIGNATURE-----




-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
Vivek Khera
2002-08-02 19:32:04 UTC
Permalink
CC> -----BEGIN PGP SIGNED MESSAGE-----
CC> Hash: SHA1

CC> Here is questionable advertising that I noticed:
CC> http://www.google.com/search?cat=&meta=&q=spamassassin


CC> A google search on spamassassin returns:
CC> - ----------====
CC> SPAMASSASSIN - Stop 98% of Spam with ActiveState
CC> www.ActiveState.com/stopspam Free evaluation. Sponsored Link

ActiveState sells a mail filtering product that incorporates
SpamAssassin. I think it is fair to identify that in their marketing.

They're definitely white-hat folks, well known in the open source
community.
--
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Vivek Khera, Ph.D. Khera Communications, Inc.
Internet: ***@kciLink.com Rockville, MD +1-240-453-8497
AIM: vivekkhera Y!: vivek_khera http://www.khera.org/~vivek/


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
Craig R.Hughes
2002-08-02 23:28:45 UTC
Permalink
They actually I believe modify SpamAssassin and add some extra
stuff in there, so it's *not* SpamAssassin that they're selling,
but rather something built on the SpamAssassin base. Yes, they
should be able to say in their marketing that they *use*
SpamAssassin, but it's not quite kosher to claim that they *are*
SpamAssassin.

On this topic, it's amazing how fast people move when you ask
them really nicely :) The "Guidelines for Use" document is in
more or less final draft form now, and I hope to have it ready
to send to the list here this evening or tomorrow morning.

C
Post by Vivek Khera
CC> -----BEGIN PGP SIGNED MESSAGE-----
CC> Hash: SHA1
CC> http://www.google.com/search?cat=&meta=&q=spamassassin
CC> - ----------====
CC> SPAMASSASSIN - Stop 98% of Spam with ActiveState
CC> www.ActiveState.com/stopspam Free evaluation. Sponsored Link
ActiveState sells a mail filtering product that incorporates
SpamAssassin. I think it is fair to identify that in their marketing.
They're definitely white-hat folks, well known in the open source
community.
-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
Malte S. Stretz
2002-08-03 12:36:30 UTC
Permalink
Post by Craig R.Hughes
CC>[...]
CC> - ----------====
CC> SPAMASSASSIN - Stop 98% of Spam with ActiveState
CC> www.ActiveState.com/stopspam Free evaluation. Sponsored Link
They actually I believe modify SpamAssassin and add some extra
stuff in there, so it's *not* SpamAssassin that they're selling,
but rather something built on the SpamAssassin base. Yes, they
should be able to say in their marketing that they *use*
SpamAssassin, but it's not quite kosher to claim that they *are*
SpamAssassin.
Here's another "Sponsored Link" from Google:
| SpamAssassin Included
| Don't download spam! Stop it on the
| server with low cost web hosting.
| www.esosoft.com

You're redirected to [1] which has a nice description of SpamAssassin's
features but doesn't make it very clear that SA isn't made by EsoSoft.
The conclusion is:
| Stop Spam at EsoSoft!
| SpamAssassin comes standard with each Basic Web Hosting account. If spam
| is ruining your Internet life, don't wait another day. Host your site at
| EsoSoft and put an end to spam!
Post by Craig R.Hughes
[...]
Malte

[1]http://www.esosoft.com/spamassassin/
--
-- Coding is art.
--
-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
Craig R.Hughes
2002-08-03 14:23:46 UTC
Permalink
Post by Malte S. Stretz
| SpamAssassin Included
| Don't download spam! Stop it on the
| server with low cost web hosting.
| www.esosoft.com
So I'd say "SpamAssassin included" is probably OK in the ad.
Post by Malte S. Stretz
You're redirected to [1] which has a nice description of SpamAssassin's
features but doesn't make it very clear that SA isn't made by EsoSoft.
| Stop Spam at EsoSoft!
| SpamAssassin comes standard with each Basic Web Hosting
account. If spam
| is ruining your Internet life, don't wait another day. Host
your site at
| EsoSoft and put an end to spam!
Again, this is saying that they're selling accounts which use
SpamAssassin -- I think that is probably OK. I haven't looked
at the site though, just your extract -- I'm heading out now,
but I'll take a look when I get back.

C



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
Matt Sergeant
2002-08-05 10:06:51 UTC
Permalink
Post by Craig R.Hughes
They actually I believe modify SpamAssassin and add some extra
stuff in there, so it's *not* SpamAssassin that they're selling,
but rather something built on the SpamAssassin base. Yes, they
should be able to say in their marketing that they *use*
SpamAssassin, but it's not quite kosher to claim that they *are*
SpamAssassin.
Interesting - I don't actually get an ActiveState link when I do that
search. Anyway, they don't claim to be SpamAssassin (at least not in the
literature I've read, or in speaking to them), but they don't mind
talking about using SpamAssassin. Either way, we should be careful not
to alienate them, as I spoke in depth with them at the conference about
how they'd like to give back to the project now - hopefully something
will come of that in talks with them in the next few weeks. There are
definitely some technical issues with how they integrate spamassassin,
but nothing showstopping, and they have their own brand which I'm sure
they're happy with.

Matt.



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
Craig R.Hughes
2002-08-05 19:28:17 UTC
Permalink
Probably cos you're in the UK and they're doing geo-targetting
with Adwords or something. Screenshot attached. We're not
trying to be hostile to ActiveState or anyone else, we just want
to make sure that they and everyone else makes it clear how/when
they're using SpamAssassin, and that if what they're selling is
modified in some way, and not "regular" SpamAssassin, that they
make that clear. The rules for this are to some extent covered
in the PAL, but we're basically trying to strengthen the PAL by
solidifying the ownership of the name.
Justin Mason
2002-08-02 10:11:28 UTC
Permalink
My intention in consolidating the trademark rights into Deersoft
was primarily to ensure that others could not call a product
"SpamAssassin" unless it really was just that -- that is noone
would be able to produce some other shitty spam filter, and the
name developed by Justin for it. However, of course, anyone
should be free to use the real, authentic SpamAssassin and let
people know that's what they're doing. Basically, if you call
SpamAssassin "SpamAssassin", that's fine, if you call something
else "SpamAssassin", that's not fine.
Also worth noting that I've been working with Deersoft on this, so it
can't be used as a Deersoft grab tactic if they ever Turn Bad ;)

--j.


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
Tony L. Svanstrom
2002-08-06 16:39:50 UTC
Permalink
My intention in consolidating the trademark rights into Deersoft was
primarily to ensure that others could not call a product "SpamAssassin"
unless it really was just that -- that is noone would be able to produce
some other shitty spam filter, and the name developed by Justin for it.
However, of course, anyone should be free to use the real, authentic
SpamAssassin and let people know that's what they're doing. Basically, if
you call SpamAssassin "SpamAssassin", that's fine, if you call something
else "SpamAssassin", that's not fine.
Also worth noting that I've been working with Deersoft on this, so it can't
be used as a Deersoft grab tactic if they ever Turn Bad ;)
IMHO it's wrong that any company is allowed to use the name of a public
project commercially, while others aren't´; esp. when it's a name that tells
Average Joe that that version is so much better than the free one.

I'm not worried about Deersoft turning bad... eh, worse... though... ;-)


/Tony
--
# Per scientiam ad libertatem! // Through knowledge towards freedom! #
# Genom kunskap mot frihet! =*= (c) 1999-2002 ***@svanstrom.com =*= #

perl -e'print$_{$_} for sort%_=`lynx -source svanstrom.com/t`'



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
Rich Wellner
2002-08-08 13:58:01 UTC
Permalink
Post by Tony L. Svanstrom
IMHO it's wrong that any company is allowed to use the name of a public
project commercially, while others aren't´; esp. when it's a name that tells
Average Joe that that version is so much better than the free one.
Dude, take that one up with Redhat, Suse, IBM or any of the other folks using
the name of a public project (i.e. Linux) for a commercial product. ;-)

This is not only a concept with precedent it's actually quite common.

rw2


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
Tony L. Svanstrom
2002-08-08 14:57:53 UTC
Permalink
Post by Rich Wellner
Post by Tony L. Svanstrom
IMHO it's wrong that any company is allowed to use the name of a public
project commercially, while others aren't´; esp. when it's a name that tells
Average Joe that that version is so much better than the free one.
Dude, take that one up with Redhat, Suse, IBM or any of the other folks using
the name of a public project (i.e. Linux) for a commercial product. ;-)
This is not only a concept with precedent it's actually quite common.
Don't make me get your moma to talk to you about jumping off cliffs just
because the other kids are doing it...


Besides... you can't really compare this with Linux, because you've got "Red
Hat Linux", "SuSE Linux", "Yellow Dog Linux", "unitedlinux" and so on, while
Deersoft won't allow other products to base their names on "SpamAssassin".
Calling their product for "SpamAssassin Pro" is, IMHO, too "marketing greedy"
to begin with; but calling it "SpamAssassin Pro" and at the same time
preventing anyone else from using names like "Blue Boot SpamAssassin"... no...
no... no... me no like...

(For the comparison to have worked it should have been something like this:
"Red Hat, a Linux-based OS", "Yellow Dog OS - A LinuxOS for PPC", "SuSE, with
Linux".)


/Tony
PS Once again I like to point out that this is a general "I don't think it
should be done that way"-comment, and that I'm not anti-Deersoft for them doing
it; I might even have done exactly the same myself if I'd been Craig.
--
# Per scientiam ad libertatem! // Through knowledge towards freedom! #
# Genom kunskap mot frihet! =*= (c) 1999-2002 ***@svanstrom.com =*= #

perl -e'print$_{$_} for sort%_=`lynx -source svanstrom.com/t`'



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
Craig R.Hughes
2002-08-08 21:56:14 UTC
Permalink
Post by Tony L. Svanstrom
(For the comparison to have worked it should have been
"Red Hat, a Linux-based OS", "Yellow Dog OS - A LinuxOS for
PPC", "SuSE, with
Linux".)
So I'm guessing there's very little legal footing for defence of
the Linux trademark, precisely because of this.
Post by Tony L. Svanstrom
/Tony
PS Once again I like to point out that this is a general "I
don't think it
should be done that way"-comment, and that I'm not
anti-Deersoft for them doing
it; I might even have done exactly the same myself if I'd been Craig.
Thanks :)

We are trying to be "good citizens" though here. Anyone who
wants to use the SpamAssassin name in a reasonable way is likely
to be able to get an extremely good deal (don't know what the
minimum legal requirement is? free?) on licensing the trademark
from us. As I understand it though, we do need to review and
approve any use of the mark which doesn't fit with the
guidelines for use, or else we lose the legal protection of the
mark, and folks like Gotts can be as abusive as they want. But
IANAL.

C



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
Rich Wellner
2002-08-08 22:14:26 UTC
Permalink
Post by Tony L. Svanstrom
"Red Hat, a Linux-based OS", "Yellow Dog OS - A LinuxOS for PPC", "SuSE, with
Linux".)
So I'm guessing there's very little legal footing for defence of the Linux
trademark, precisely because of this.
The odds of you asking your lawyers about this are far greater than the odds
that I've asked mine (since I *know* that I haven't), but my understanding
was...

That they use the name isn't a problem. Where you to give me permission, I
could come out with a line of beer mugs named "SpamAssassin" and be just fine.
The key is that you, as the owner of the mark, give me permission, not that
the name is adorned with qualifiers

Linus does protect the name. That he grants a lot of people the right to use
it doesn't change that he protects it. There was a slashdot about this a
while back...

http://slashdot.org/articles/00/01/19/0828245.shtml

rw2


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
Craig R.Hughes
2002-08-09 16:00:28 UTC
Permalink
Post by Rich Wellner
That they use the name isn't a problem. Where you to give me
permission, I
could come out with a line of beer mugs named "SpamAssassin"
and be just fine.
The key is that you, as the owner of the mark, give me
permission, not that
the name is adorned with qualifiers
Actually, you could produce a line of SpamAssassin beer mugs
even without permission -- the trademark does not apply to beer
mugs, just to spam-filtering software. So you can't make a new
spam filter and call it SpamAssassin.
Post by Rich Wellner
Linus does protect the name. That he grants a lot of people
the right to use
it doesn't change that he protects it. There was a slashdot
about this a
while back...
I'll take a look at that. We are of course searching for models
which can make everyone happy while still allowing us to defend
the name from real abuse.

C



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
Tony L. Svanstrom
2002-08-08 23:26:20 UTC
Permalink
We are trying to be "good citizens" though here. Anyone who wants to use the
SpamAssassin name in a reasonable way is likely to be able to get an
extremely good deal (don't know what the minimum legal requirement is?
free?) on licensing the trademark from us.
I'm a Swede targeting the scandinavian market, incl. the englishspeaking
ones... you got "SpamAssassin" protected in Sweden? ;-)


/Tony
--
# Per scientiam ad libertatem! // Through knowledge towards freedom! #
# Genom kunskap mot frihet! =*= (c) 1999-2002 ***@svanstrom.com =*= #

perl -e'print$_{$_} for sort%_=`lynx -source svanstrom.com/t`'



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
Craig R.Hughes
2002-08-09 16:02:38 UTC
Permalink
Post by Tony L. Svanstrom
Post by Craig R.Hughes
We are trying to be "good citizens" though here. Anyone who
wants to use the
SpamAssassin name in a reasonable way is likely to be able to get an
extremely good deal (don't know what the minimum legal requirement is?
free?) on licensing the trademark from us.
I'm a Swede targeting the scandinavian market, incl. the
englishspeaking
ones... you got "SpamAssassin" protected in Sweden? ;-)
Yes.

C



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
Justin Mason
2002-08-06 17:19:29 UTC
Permalink
Excellent. After that, I'd suggest we work on stabilizing (ie
no new features) for a bit, and getting 2.40 released. I think
we're in reasonably good shape now with 2.40 (plus or minus some
build issues occasionally creeping in), and Razor2 is starting
to get a little less flakey (minus one or two server issues in
the last few days).
Yeah, I agree. Although note: I'd like to see the proposal to use
Received IPs in the AWL, get into 2.40 too. That's being actively
exploited right now.

Razor2 is definitely getting solid, and getting some good hits too!

--j.


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
Craig R.Hughes
2002-08-06 21:45:23 UTC
Permalink
Post by Justin Mason
Excellent. After that, I'd suggest we work on stabilizing (ie
no new features) for a bit, and getting 2.40 released. I think
we're in reasonably good shape now with 2.40 (plus or minus some
build issues occasionally creeping in), and Razor2 is starting
to get a little less flakey (minus one or two server issues in
the last few days).
Yeah, I agree. Although note: I'd like to see the proposal to use
Received IPs in the AWL, get into 2.40 too. That's being actively
exploited right now.
Yeah, AWL IPs is a good thing for 2.40
Post by Justin Mason
Razor2 is definitely getting solid, and getting some good hits too!
When it works, yeah. But it's so flakey I have to turn it off
so that my mail can be guaranteed to actually get through in a
decent timeframe.

C



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
Justin Mason
2002-08-21 10:58:48 UTC
Permalink
body CALL_NOW /\bCALL NOW/i
IMO it probably started life without the /i, and some well-meaning soul
added it. I really don't like it -- it's a very common piece of text....

--j.


-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Michael Moncur
2002-08-21 11:29:04 UTC
Permalink
Post by Justin Mason
body CALL_NOW /\bCALL NOW/i
IMO it probably started life without the /i, and some well-meaning soul
added it. I really don't like it -- it's a very common piece of text....
I don't like it either, it's the kind of rule that will test great on the
corpus, get a high score, and then lurk in the background, waiting to pounce
on anyone who says "I'll give you a call now that I'm back from vacation."

Not to mention the National Organization for Women.

--
Michael Moncur mgm at starlingtech.com http://www.starlingtech.com/
"The cure for writer's cramp is writer's block." --Inigo DeLeon



-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Scott A Crosby
2002-08-21 11:45:24 UTC
Permalink
Post by Justin Mason
body CALL_NOW /\bCALL NOW/i
IMO it probably started life without the /i, and some well-meaning soul
added it. I really don't like it -- it's a very common piece of text....
How about 2 versions of it? One thats case-sensitive, the other thats not?


/\bCALL NOW/
/\bcall now/i

Where one gets a big score, the other a small score.

Scott


-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Justin Mason
2002-08-21 11:55:15 UTC
Permalink
Post by Scott A Crosby
How about 2 versions of it? One thats case-sensitive, the other thats not?
/\bCALL NOW/
/\bcall now/i
Where one gets a big score, the other a small score.
I like the top one. The problem I have with the bottom one is that we
just don't have a lot of mails in the corpus talking about, as Mike M
said, the National Organisation of Women or similar, or talking about
people telephoning other people, so the GA doesn't really know that it's
quite a common formulation in the english language.

I mean, if your boss mails you saying "we need to make a call NOW",
that's a perfectly valid nonspam formulation -- but it's not in our
corpora very much I'd imagine, so the GA will give it too high a score.

--j.


-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Craig R.Hughes
2002-08-22 16:15:26 UTC
Permalink
Aside from anything else, /\bword1 word2\b/i is always going to
be covered by the spam phrases checking.

/\bWORD1 WORD2\b/ though is a special case. Let's go ahead and
drop the case-insensitive bit.

C
Post by Justin Mason
Post by Scott A Crosby
How about 2 versions of it? One thats case-sensitive, the
other thats not?
/\bCALL NOW/
/\bcall now/i
Where one gets a big score, the other a small score.
I like the top one. The problem I have with the bottom one is that we
just don't have a lot of mails in the corpus talking about, as Mike M
said, the National Organisation of Women or similar, or talking about
people telephoning other people, so the GA doesn't really know
that it's
quite a common formulation in the english language.
I mean, if your boss mails you saying "we need to make a call NOW",
that's a perfectly valid nonspam formulation -- but it's not in our
corpora very much I'd imagine, so the GA will give it too high a score.
--j.
-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
_______________________________________________
Spamassassin-talk mailing list
https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Justin Mason
2002-08-21 12:34:26 UTC
Permalink
In essence, the scheme sounds like putting a legal (rather than
cryptographic) certificate in email, then klomping down on the those
who forge the certificate.
exactly. I think, as long as Habeas have a huge stack of lawyers and
funds to keep 'em chugging along, it'll work. I'm not convinced it's
_quite_ that easy to do that though ;) -- but I'll wish them good luck
all the same, and I'm happy to put the HABEAS_SWE test in SpamAssassin to
help out.
I'm not saying that certs are a bad idea.. But I don't want to have
yet more time and money spent on lawyers, so how about cryptographic
certs?
BTW, this is another place where we can help out. Already SpamAssassin
marks PGP certs with a compensation test, and we should probably do the
same for S/MIME signed messages -- so be sure to put it about that signing
mail is Good Against Spam in a similar way (to a certain extent).

(I presume BTW that you can see why Habeas didn't use a cryptographic
signature in their algorithm? ie. no legal protection under current
world-wide law?)

--j.


-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Scott A Crosby
2002-08-21 13:09:05 UTC
Permalink
Post by Justin Mason
In essence, the scheme sounds like putting a legal (rather than
cryptographic) certificate in email, then klomping down on the those
who forge the certificate.
exactly. I think, as long as Habeas have a huge stack of lawyers and
funds to keep 'em chugging along, it'll work. I'm not convinced it's
Cryptographic certs would work just as well.
Post by Justin Mason
(I presume BTW that you can see why Habeas didn't use a cryptographic
signature in their algorithm? ie. no legal protection under current
world-wide law?)
Not at all....

A cryptographic cert is unforgable. Thats like passing a law that I
can't put your PGP signature on my messages... Well DUH, thats
pointless. The whole concept of cryptographic certs is that I couldn't
forge yours even if I wanted.

So, you *don't* need legal protection to prevent undesirables from
signing with your cryptographic cert..

Just a key put into a global keyring of, say, ""this isn't spam""[1]

Sign your messages with that key.

The cryptographic cert has *authenticity*. While it may be true that
such authenticity is considered insufficient weight for contract or
other law, that isn't relevant to this issue.

IE, the cert in this case just indicates that its not spam.. The cert
may not have sufficient legal weight to make the email a binding
contract worldwide, but thats unneeded in this scenario.

Perhaps you confused this issue?

Scott


[1] Not quite that simple, because whomever controls that keyring gets
to abuse their power, to use collateral damage as a weapon against any
they wish... Unacceptable.


-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Tony Hoyle
2002-08-21 12:39:46 UTC
Permalink
Post by Vince Puzzella
-----Original Message-----
Sent: 21 August 2002 07:39
To: Matthew Cline
Cc: SpamAssassin Talk ML
Subject: Re: [SAtalk] SA In The News
talking with Dan Kohn about this concept for a while now (since
the time I was contemplating Deersoft), and I think it sounds
like quite a promising concept -- we'll see how well they do at
enforcing their copyright against chinese spammers and the like.
I don't see how it can work, really. Such patents are (AFAIK) only
valid in the US & (possibly) contries that have agreements with the US..
If a spammer from China decides to use it then they can't be touched
legally, so the method rapidly becomes useless.

Tony


-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Craig R.Hughes
2002-08-22 16:39:01 UTC
Permalink
The patent covers the actions of would-be anti-spammers, not
spammers. The copyright and its license prevent spammers from
using the Haiku, and most countries in the world share copyright
agreements. The trademark is probably harder to enforce
overseas, but I think that's really a third crutch which
provides added firepower against US-based spammers.

Basically, the copyright is the tool used against spammers
The patent is to protect Habeas from other people copying their
business idea
The trademark is extra ammo against spammers who qualify.

C
Post by Tony Hoyle
Post by Vince Puzzella
talking with Dan Kohn about this concept for a while now (since
the time I was contemplating Deersoft), and I think it sounds
like quite a promising concept -- we'll see how well they do at
enforcing their copyright against chinese spammers and the like.
I don't see how it can work, really. Such patents are (AFAIK) only
valid in the US & (possibly) contries that have agreements with the US..
If a spammer from China decides to use it then they can't be touched
legally, so the method rapidly becomes useless.
-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Justin Mason
2002-08-21 13:32:32 UTC
Permalink
Post by Scott A Crosby
Just a key put into a global keyring of, say, ""this isn't spam""[1]
Sign your messages with that key.
The cryptographic cert has *authenticity*. While it may be true that
such authenticity is considered insufficient weight for contract or
other law, that isn't relevant to this issue.
IE, the cert in this case just indicates that its not spam.. The cert
may not have sufficient legal weight to make the email a binding
contract worldwide, but thats unneeded in this scenario.
OK -- let's see. So the user generates a key, and asks a global keyring
maintainer to add its sig to the global "this isn't spam" keyring. Then
the user sends spam. what happens then? Does the global keyring
maintainer revoke the key? In that case, where is this effectively
different from an DNSBL?

DNSBL = "this originating IP is listed in the DNSBL"
keyring = "this originating key is listed as 'revoked'"

Not quite the same thing as what Habeas are proposing.

Or do you envisage that the global keyring maintainer can then sue the
spammer for breach of contract?

I think using traditional methods (copyright etc.) that are going to be
well-supported by worldwide legal systems and worldwide judiciary ;) is a
better way to make sure the 'breach of contract' suing works.

--j.


-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Justin Mason
2002-08-22 16:36:48 UTC
Permalink
I thought we'd sort of decided that after 2.40 we'd switch the
meaning of the "lang" prefix to actually mean "run these rules
if language matches", where "language" is what TextCat spits out
for a message, as opposed to using the locale which is how
things currently work.
ah, right -- I'd forgotten that. :(

doesn't help with the finding-broken-rules QA case though...

--j.
--
'Justin Mason' => { url => http://jmason.org/ , blog => http://taint.org/ }


-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Craig R.Hughes
2002-08-22 18:17:14 UTC
Permalink
Post by Justin Mason
I thought we'd sort of decided that after 2.40 we'd switch the
meaning of the "lang" prefix to actually mean "run these rules
if language matches", where "language" is what TextCat spits out
for a message, as opposed to using the locale which is how
things currently work.
ah, right -- I'd forgotten that. :(
doesn't help with the finding-broken-rules QA case though...
It does, if you have a corpus of german spam and nonspam to feed
it, and you're not in a de locale.

C



-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Justin Mason
2002-08-22 17:58:01 UTC
Permalink
Are you filtering the nonspamtrap for spam when those
newsletters sold your address to someone you didn't sign up
with? You should probably manually verify that all the mail in
the trap is in fact nonspam!
Yep. so far it's all opt-in stuff, no "affiliate deals" etc.

--j.


-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Justin Mason
2002-08-23 11:27:58 UTC
Permalink
I'm a bit confused by them -- don't think I've ever seen
anything like that myself. One possibility is an SO_LINGER
problem or something, but your rate is low enough that I doubt
the problem lies there. Might be worth checking though with
netstat to see if there are any lingering sockets lying around.
Kelsey -- you haven't changed anything at the OS level, like libc or perl?

--j.


-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Kelsey Cummings
2002-08-23 17:51:30 UTC
Permalink
Post by Justin Mason
I'm a bit confused by them -- don't think I've ever seen
anything like that myself. One possibility is an SO_LINGER
problem or something, but your rate is low enough that I doubt
the problem lies there. Might be worth checking though with
netstat to see if there are any lingering sockets lying around.
Kelsey -- you haven't changed anything at the OS level, like libc or perl?
naw. stock 6.2
--
Kelsey Cummings - ***@sonic.net sonic.net
System Administrator 2260 Apollo Way
707.522.1000 (Voice) Santa Rosa, CA 95407
707.547.2199 (Fax) http://www.sonic.net/
Fingerprint = 7F 59 43 1B 44 8A 0D 57 91 08 73 73 7A 48 90 C5


-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Justin Mason
2002-08-23 14:24:23 UTC
Permalink
These are more correct -- Craig noted that I had some duplicated
log lines :(


OVERALL% SPAM% NONSPAM% S/O SCORE NAME
242606 77530 165076 0.32 0.00 (all messages)
100.000 31.957 68.043 0.32 0.00 (all messages as %)
3.643 11.399 0.000 1.00 4.34 MIME_ODD_CASE
3.650 11.416 0.002 1.00 4.39 INVALID_DATE_TZ_ABSURD
2.027 6.341 0.001 1.00 1.00 FROM_HAS_MIXED_NUMS2
1.690 5.284 0.002 1.00 1.00 DATE_YEAR_ZERO_FIRST
0.764 2.391 0.000 1.00 3.40 KOREAN_UCE_SUBJECT
0.690 2.160 0.000 1.00 4.14 X_ENC_PRESENT
0.667 2.087 0.000 1.00 4.55 X_PRECEDENCE_REF
0.652 2.041 0.000 1.00 3.46 MSGID_SPAMSIGN_1
2.483 0.001 3.649 0.00 -1.00 USER_AGENT_MUTT
0.474 1.482 0.000 1.00 1.00 COMPETE
0.375 1.172 0.000 1.00 1.00 FREE_INVESTMENT
2.587 8.079 0.008 1.00 4.30 SUBJ_FULL_OF_8BITS
0.334 1.046 0.000 1.00 2.67 TAKE_ACTION_NOW
0.323 1.010 0.000 1.00 1.00 REFINANCE_YOUR_HOME
0.322 1.009 0.000 1.00 1.00 EASY_TERMS
0.314 0.982 0.000 1.00 1.00 REVERSE_AGING
0.279 0.875 0.000 1.00 2.19 X_MAIL_ID_PRESENT
0.272 0.851 0.000 1.00 3.76 X_SERV_HOST_PRESENT
0.228 0.713 0.000 1.00 1.00 HIDDEN_ASSETS
0.211 0.660 0.000 1.00 1.00 USERNAME_IN_SUBJECT
0.211 0.660 0.000 1.00 1.50 USERNAME_IN_SUBJECT_4
0.211 0.660 0.000 1.00 1.00 USERNAME_IN_SUBJECT_5
0.211 0.660 0.000 1.00 0.50 USERNAME_IN_SUBJECT_6
0.211 0.660 0.000 1.00 0.50 USERNAME_IN_SUBJECT_2
0.211 0.660 0.000 1.00 2.00 USERNAME_IN_SUBJECT_3
2.187 0.003 3.212 0.00 -1.00 USER_AGENT_KMAIL
0.204 0.640 0.000 1.00 1.00 WRINKLES
0.198 0.619 0.000 1.00 1.51 X_X_PRESENT
0.192 0.600 0.000 1.00 4.34 CBYI
0.183 0.573 0.000 1.00 3.19 X_LIST_UNSUBSCRIBE
0.181 0.568 0.000 1.00 1.00 NOT_MLM
0.178 0.556 0.000 1.00 1.00 DIG_UP_INFO
0.316 0.000 0.464 0.00 -3.13 PGP_SIGNATURE_2
0.390 1.218 0.001 1.00 1.00 FREE_PASSWORD
0.136 0.427 0.000 1.00 2.69 HR_3113
0.134 0.418 0.000 1.00 4.52 NONEXISTENT_CHARSET
1.087 3.395 0.004 1.00 1.00 EXCUSE_FUTURE
0.126 0.395 0.000 1.00 3.65 FAKED_IP_IN_RCVD
0.266 0.000 0.391 0.00 -1.00 SIGNATURE_LONG_DENSE
0.122 0.382 0.000 1.00 1.00 COMPLAIN_TO
0.120 0.377 0.000 1.00 2.62 UCE_MAIL_ACT
0.527 1.646 0.001 1.00 1.00 FINANCIAL
0.117 0.366 0.000 1.00 3.52 FROM_BTAMAIL
0.108 0.339 0.000 1.00 3.80 WE_HATE_SPAM
0.504 1.574 0.001 1.00 4.18 FREE_CONSULTATION
0.107 0.334 0.000 1.00 1.00 MEGA_SITE
0.332 1.038 0.001 1.00 1.00 FREE_GRANT
0.326 1.018 0.001 1.00 3.41 COPY_ACCURATELY
0.208 0.000 0.306 0.00 -5.00 PATCH_UNIFIED_DIFF
0.094 0.295 0.000 1.00 4.25 NEW_DOMAIN_EXTENSIONS
0.094 0.295 0.000 1.00 4.46 PRODUCED_AND_SENT_OUT
0.092 0.289 0.000 1.00 4.20 GENTLE_FEROCITY
0.090 0.282 0.000 1.00 4.85 FORGED_GW05_RCVD
0.089 0.279 0.000 1.00 1.00 INCOME
0.310 0.970 0.001 1.00 1.00 HTML_COMMENT_8BITS
0.084 0.264 0.000 1.00 1.00 MARKUP_RAND
0.084 0.263 0.000 1.00 4.28 VJESTIKA
0.082 0.258 0.000 1.00 1.00 MIME_BOUND_OPTIN
0.082 0.255 0.000 1.00 2.27 X_STORMPOST_TO
0.082 0.255 0.000 1.00 2.00 RATWARE_STORM
0.073 0.230 0.000 1.00 3.03 LOTS_OF_CC_LINES
0.070 0.221 0.000 1.00 2.00 RATWARE_VC_IPA
0.070 0.218 0.000 1.00 1.00 VIAGRA_COMBO
0.066 0.208 0.000 1.00 1.00 TO_RECIP_MARKER
0.066 0.206 0.000 1.00 1.00 PORN_MEMBERSHIP
0.063 0.196 0.000 1.00 1.80 EXCUSE_5
1.080 3.370 0.005 1.00 1.00 CONSOLIDATE_DEBT
0.059 0.183 0.000 1.00 4.46 ADDRESSES_ON_CD
0.057 0.179 0.000 1.00 1.00 NO_INVENTORY
0.054 0.170 0.000 1.00 1.00 HR_4176
0.246 0.767 0.001 1.00 4.58 FORGED_EUDORAMAIL_RCVD
0.050 0.157 0.000 1.00 3.00 RATWARE_OE_PI
0.502 1.567 0.002 1.00 1.18 SENT_IN_COMPLIANCE
0.103 0.000 0.152 0.00 -2.00 BUGZILLA_BUG
0.102 0.000 0.150 0.00 -10.00 GENUINE_EBAY_RCVD
0.046 0.144 0.000 1.00 1.00 CABLE_CONVERTER
0.046 0.144 0.000 1.00 1.93 PORN_6
0.237 0.740 0.001 1.00 2.90 EXCUSE_4
0.046 0.143 0.000 1.00 1.00 HARDCORE_PORN
0.045 0.142 0.000 1.00 4.13 PURE_PROFIT
0.044 0.138 0.000 1.00 2.00 RATWARE_MMAILER
0.043 0.135 0.000 1.00 2.00 RATWARE_EVAMAIL
0.796 2.483 0.004 1.00 1.00 REFINANCE
0.477 1.490 0.002 1.00 2.50 SPAM_PHRASE_34_55
0.040 0.126 0.000 1.00 2.00 RATWARE_SCREWUP_1
0.040 0.125 0.000 1.00 2.73 SERIOUS_ONLY
0.038 0.120 0.000 1.00 2.00 RATWARE_JIXING
0.351 1.095 0.001 1.00 1.00 BEEN_TURNED_DOWN
0.037 0.116 0.000 1.00 2.94 PENNIES_A_DAY
0.212 0.663 0.001 1.00 4.28 WE_HONOR_ALL
0.331 1.034 0.001 1.00 0.39 CLICK_TO_REMOVE_2
0.031 0.095 0.000 1.00 3.66 MICRO_CAP_WARNING
0.031 0.095 0.000 1.00 2.10 LONG_NUMERIC_HTTP_ADDR
0.029 0.092 0.000 1.00 1.00 MIME_BOUND_HASHES
0.059 0.000 0.087 0.00 -1.00 USER_AGENT_IMP
0.027 0.084 0.000 1.00 1.00 CREDITORS_CALLING
0.026 0.080 0.000 1.00 2.00 RATWARE_PASCUAL
0.190 0.595 0.001 1.00 2.61 JODY
0.024 0.076 0.000 1.00 1.00 PORN_PASSWORD
0.024 0.075 0.000 1.00 1.41 S_1618
0.023 0.074 0.000 1.00 1.00 ANOTHER_NET_AD
0.023 0.071 0.000 1.00 2.56 MDAEMON_2_7_4
0.399 1.246 0.002 1.00 1.00 HGH
0.021 0.064 0.000 1.00 3.21 SPAM_FORM_RETURN
0.021 0.064 0.000 1.00 2.00 RATWARE_CURMAIL
22.025 0.101 32.321 0.00 -3.38 IN_REP_TO
0.020 0.062 0.000 1.00 4.00 VAR_REF_IN_RECEIVED
0.658 2.052 0.004 1.00 3.46 FAKED_UNDISC_RECIPS
0.018 0.055 0.000 1.00 2.00 RATWARE_XMAILER
0.018 0.055 0.000 1.00 2.00 RATWARE_EPAPER
2.254 7.017 0.017 1.00 2.00 HEADER_8BITS
2.687 8.365 0.021 1.00 1.00 CLICK_HERE_CAPS_LINK
0.016 0.052 0.000 1.00 1.00 ONLINE_BIZ_OPS
0.016 0.052 0.000 1.00 2.09 WWW_CLIK4YOU_COM
0.016 0.049 0.000 1.00 1.00 GREEN_EXCUSE_2
0.015 0.046 0.000 1.00 1.00 IN_ACCORDANCE_WITH_LAWS
0.014 0.044 0.000 1.00 1.00 POPLAUNCH
0.014 0.043 0.000 1.00 1.00 GREEN_EXCUSE_1
0.014 0.043 0.000 1.00 3.20 UNNEEDED_HTML_ENCODING
0.014 0.043 0.000 1.00 1.00 EXCUSE_18
0.014 0.043 0.000 1.00 1.00 EXCUSE_8
0.028 0.000 0.041 0.00 -4.00 Q_FOR_SELLER
0.156 0.486 0.001 1.00 1.00 VIAGRA_ONLINE
0.013 0.040 0.000 1.00 2.00 RATWARE_YAM
0.013 0.040 0.000 1.00 1.00 MIME_BOUND_MAIL_BOUND
0.012 0.037 0.000 1.00 1.00 FREE_LEADS
4.917 15.287 0.046 1.00 4.04 SUBJ_HAS_SPACES
0.012 0.036 0.000 1.00 2.00 RATWARE_POWERC
0.151 0.471 0.001 1.00 3.52 RESISTANCE_IS_FUTILE
0.011 0.035 0.000 1.00 1.00 FREEWEBCO_NET_URL
0.011 0.035 0.000 1.00 1.00 NO_GIMMICK
0.011 0.034 0.000 1.00 0.22 SPAM_FORM
0.011 0.034 0.000 1.00 1.00 MIME_BOUND_MIME_BOUND
0.022 0.000 0.032 0.00 -5.00 PATCH_CONTEXT_DIFF
0.010 0.031 0.000 1.00 1.00 BUY_JUDGEMENTS
0.243 0.758 0.001 1.00 2.19 PARA_A_2_C_OF_1618
0.009 0.030 0.000 1.00 1.00 YELLOWSUN
0.009 0.027 0.000 1.00 2.00 RATWARE_DIFFOND
0.009 0.027 0.000 1.00 1.00 MARKUP_SSPL
0.008 0.026 0.000 1.00 2.00 RATWARE_HSU
0.017 0.000 0.025 0.00 -20.00 HABEAS_SWE
0.017 0.000 0.025 0.00 -4.00 MAILMAN_CONFIRM
0.136 0.426 0.001 1.00 3.50 PREST_NON_ACCREDITED
0.008 0.025 0.000 1.00 1.00 INTL_EXEC_GUILD
0.008 0.025 0.000 1.00 1.00 FREEWEBHOSTINGCENTRAL
0.016 0.000 0.023 0.00 -2.00 CRON_ENV
0.007 0.021 0.000 1.00 1.00 NO_MEDICAL
0.007 0.021 0.000 1.00 1.00 FILTERED_BY_WORLDREMOVE
0.007 0.021 0.000 1.00 2.00 RATWARE_SEEDNET
0.006 0.019 0.000 1.00 1.00 MIME_BOUND_SEP1
0.006 0.019 0.000 1.00 1.00 MIME_BOUND_HEX14
0.006 0.018 0.000 1.00 1.00 MURKOWSKI_CRUFT
0.006 0.018 0.000 1.00 2.00 RATWARE_CHARSET
0.006 0.018 0.000 1.00 2.00 RATWARE_CARETOP
0.006 0.018 0.000 1.00 1.00 NATURAL_VIAGRA
0.006 0.018 0.000 1.00 0.00 TO_INVESTORS
0.005 0.017 0.000 1.00 2.67 CORRUPT_MSGID
0.005 0.017 0.000 1.00 1.00 WWW_AUTOREMOVE_COM
0.011 0.000 0.016 0.00 -3.00 MSN_GROUPS
0.011 0.000 0.016 0.00 1.00 SUBJECT_IS_IN_REVIEW
0.208 0.649 0.001 1.00 1.00 SUBJ_GUARANTEED
0.005 0.014 0.000 1.00 1.00 PRIORITY_MAIL
0.009 0.000 0.014 0.00 -1.00 USER_AGENT_TONLINE
0.352 1.096 0.002 1.00 4.61 BILL_1618
0.004 0.013 0.000 1.00 0.20 JUST_MAILED_PAGE
14.883 0.081 21.835 0.00 -0.10 REFERENCES
0.004 0.012 0.000 1.00 1.00 T_SALE
0.004 0.012 0.000 1.00 2.66 POST_IN_RCVD
0.004 0.012 0.000 1.00 1.00 SPAM_FORM_INPUT
0.004 0.012 0.000 1.00 1.00 INTERNET_TERROR_RANT
0.004 0.012 0.000 1.00 1.00 NO_MIDDLEMAN
0.004 0.012 0.000 1.00 1.00 NO_SELLING
0.003 0.010 0.000 1.00 2.00 RATWARE_LC_OUTLOOK
0.003 0.010 0.000 1.00 1.00 PORN_GALLERIES
0.003 0.009 0.000 1.00 1.00 FREEMEGS_URL
0.003 0.009 0.000 1.00 3.50 CYBER_FIRE_POWER
0.002 0.008 0.000 1.00 4.25 SHORT_RECEIVED_LINE
0.002 0.008 0.000 1.00 2.00 RATWARE_MATCHMAKER
0.002 0.008 0.000 1.00 1.00 DISCONTINUE
0.002 0.008 0.000 1.00 2.00 RATWARE_EBIZ
0.002 0.008 0.000 1.00 2.00 RATWARE_OPTIN
0.103 0.322 0.001 1.00 1.00 APPLY_ONLINE
0.005 0.000 0.007 0.00 -10.00 EVITE
0.566 1.761 0.005 1.00 1.00 LOW_PAYMENT
0.002 0.006 0.000 1.00 1.00 EXCUSE_ES_01
0.002 0.006 0.000 1.00 1.00 REMOVE_ES_01
0.002 0.006 0.000 1.00 1.00 NIGERIAN_SCAM_15
0.002 0.006 0.000 1.00 2.00 RATWARE_CSMTP
0.002 0.006 0.000 1.00 1.00 NUMERIC_COMMENT
0.002 0.005 0.000 1.00 1.00 NIGERIAN_SCAM_12
0.002 0.005 0.000 1.00 2.00 RATWARE_CBLAST
0.002 0.005 0.000 1.00 -0.50 FROM_NEWS_LIST
0.002 0.005 0.000 1.00 1.00 RATWARE_V3161
0.002 0.005 0.000 1.00 1.00 NIGERIAN_SCAM_2
0.002 0.005 0.000 1.00 3.12 CHARSET_FARAWAY_HEADERS
0.002 0.005 0.000 1.00 1.00 OUTSTANDING_VALUE
0.094 0.294 0.001 1.00 2.96 STOCK_ALERT
0.473 1.470 0.004 1.00 2.82 SECTION_301
0.686 2.132 0.007 1.00 1.00 FREE_PORN
0.001 0.004 0.000 1.00 1.00 NIGERIAN_SCAM_7
0.001 0.004 0.000 1.00 1.00 SPAM_FORM_ACTION
0.001 0.004 0.000 1.00 1.00 NIGERIAN_SCAM_4
0.001 0.004 0.000 1.00 1.00 NIGERIAN_SCAM_5
0.001 0.004 0.000 1.00 1.00 WWW_DIRECTFORCEMARKETING_COM
0.001 0.004 0.000 1.00 1.00 T_LOW_PRICE
0.001 0.004 0.000 1.00 1.00 NIGERIAN_SCAM_14
0.901 2.799 0.010 1.00 4.35 MORTGAGE_RATES
0.001 0.003 0.000 1.00 1.00 RATWARE_11
0.001 0.003 0.000 1.00 1.00 NIGERIAN_SCAM_6
0.001 0.003 0.000 1.00 1.00 PSYCHIC
0.001 0.003 0.000 1.00 1.00 OBFUSCATING_JAVASCRIPT
0.001 0.003 0.000 1.00 1.00 T_GET_PAID
0.001 0.003 0.000 1.00 1.00 RATWARE_40
0.001 0.003 0.000 1.00 1.00 EU_200_32_CE
0.001 0.003 0.000 1.00 3.70 COPY_DVDS
0.001 0.003 0.000 1.00 1.00 E_WEBHOSTCENTRAL_URL
0.001 0.003 0.000 1.00 1.00 THIS_AINT_JUNK
0.001 0.003 0.000 1.00 1.00 NIGERIAN_SCAM_10
0.001 0.003 0.000 1.00 1.00 FROM_TOPICA
0.001 0.003 0.000 1.00 1.00 NIGERIAN_SCAM
0.142 0.441 0.001 1.00 3.08 HTTP_CTRL_CHARS_HOST
2.211 6.858 0.028 1.00 3.33 TRACKER_ID
0.137 0.427 0.001 1.00 1.00 EARNINGS
1.230 3.815 0.015 1.00 4.92 MAILTO_WITH_SUBJ_REMOVE
0.000 0.001 0.000 1.00 2.16 CHARSET_FARAWAY
0.000 0.001 0.000 1.00 1.00 WWW_NETSITESFORFREE_NET
0.000 0.001 0.000 1.00 1.00 RATWARE_42
0.000 0.001 0.000 1.00 1.00 INCREDIBLE
0.000 0.001 0.000 1.00 4.20 YR_MEMBERSHIP_EXCH
0.000 0.001 0.000 1.00 1.00 BREAKTHROUGH_2
0.000 0.001 0.000 1.00 2.00 RATWARE_NETMAILER
0.000 0.001 0.000 1.00 1.00 X_FIX_PRESENT
0.000 0.001 0.000 1.00 1.00 REMOVE_ES_03
0.000 0.001 0.000 1.00 1.00 WEB4PORNO_URL
0.134 0.415 0.001 1.00 1.00 NASTY_GIRLS
1.178 0.006 1.728 0.00 -1.50 SIGNATURE_SHORT_SPARSE
0.071 0.221 0.001 1.00 3.84 KIFF
0.393 1.219 0.005 1.00 1.00 CREDIT_CARD
0.063 0.197 0.001 1.00 1.00 RATWARE_OE_MALFORMED
0.560 1.736 0.007 1.00 1.00 THE_BEST_RATE
0.115 0.356 0.001 1.00 1.00 ONLINE_PHARMACY
0.252 0.001 0.370 0.00 -1.00 DISCLAIMER_LEGALESE
0.000 0.000 0.000 0.00 -1.00 MSN_FOOTER2
0.000 0.000 0.000 0.00 -1.00 USER_AGENT_GNUS_UA
0.000 0.000 0.000 0.00 1.00 T_SUBJ_ISO885915
0.000 0.000 0.000 0.00 1.00 T_FROM_ISO885915
0.000 0.000 0.000 0.00 -10.00 NMS_CGI_NOT_BUGGY
3.837 0.028 5.625 0.01 -3.00 FROM_EGROUPS
0.153 0.473 0.002 1.00 1.00 COMPARE_RATES
1.706 5.281 0.027 0.99 4.00 FORGED_AOL_RCVD
0.320 0.992 0.004 1.00 1.00 WHILE_YOU_SLEEP
0.104 0.322 0.001 1.00 1.14 NOT_INTENDED
0.103 0.321 0.001 1.00 1.00 NO_CREDIT_CHECK
0.054 0.169 0.001 1.00 2.00 READ_TO_END
0.460 1.427 0.007 1.00 1.00 INITIAL_INVEST
0.099 0.308 0.001 1.00 1.00 FREE_INSTALL
6.137 18.958 0.115 0.99 2.20 REMOVE_PAGE
0.051 0.159 0.001 1.00 1.00 MEET_SINGLES
1.504 4.649 0.027 0.99 1.00 HTML_COMMENT_SAVED_URL
1.214 3.755 0.021 0.99 1.00 MARKETING_PARTNERS
0.415 1.285 0.007 0.99 3.24 FORM_W_MAILTO_ACTION
0.237 0.735 0.004 1.00 1.00 PAY_SITE
0.364 0.003 0.534 0.00 -6.00 EGP_HTML_BANNER
0.045 0.139 0.001 1.00 1.00 UNCLAIMED_MONEY
0.045 0.138 0.001 1.00 1.00 MICROSOFT
1.429 4.415 0.027 0.99 3.31 DATE_IN_FUTURE_12_24
0.043 0.133 0.001 1.00 2.00 RATWARE_SCREWUP_2
0.114 0.353 0.002 0.99 1.00 SAVE_ON_INSURANCE
0.041 0.128 0.001 1.00 3.09 X_PMFLAGS_PRESENT
0.041 0.128 0.001 1.00 1.00 CREDIT_BUREAU
0.041 0.126 0.001 1.00 2.00 RATWARE_GR
0.453 1.401 0.008 0.99 0.59 GUARANTEED_100_PERCENT
5.243 16.154 0.119 0.99 2.54 SUBJ_HAS_UNIQ_ID
0.232 0.717 0.004 0.99 1.00 BE_BOSS
0.322 0.996 0.006 0.99 1.00 MUST_BE_18
4.251 0.044 6.227 0.01 -1.00 USER_AGENT_PINE
0.037 0.115 0.001 0.99 3.07 EXCUSE_2
1.187 0.012 1.739 0.01 -1.00 USER_AGENT_MOZILLA_UA
0.101 0.311 0.002 0.99 1.00 CUM_SHOT
4.299 0.046 6.296 0.01 -1.00 GROUPS_YAHOO_1
0.153 0.001 0.225 0.01 -1.00 HOTMAIL_FOOTER3
0.493 1.521 0.010 0.99 1.00 LOW_INTEREST
0.097 0.299 0.002 0.99 1.00 CELEBRITY_PORN
0.035 0.107 0.001 0.99 1.00 STOP_SNORING
0.530 1.634 0.012 0.99 4.73 CALL_NOW
0.288 0.889 0.006 0.99 4.72 ADVERT_CODE
0.480 1.479 0.011 0.99 1.00 SAVE_THOUSANDS
0.172 0.530 0.004 0.99 4.56 PRINT_FORM_SIGNATURE
0.276 0.851 0.006 0.99 0.63 AOL_USERS_LINK
0.748 2.304 0.018 0.99 2.23 AS_SEEN_ON
0.061 0.190 0.001 0.99 1.00 WE_PROMISE_YOU
7.191 22.086 0.195 0.99 1.00 MISSING_MIMEOLE
0.449 1.383 0.010 0.99 2.22 HTTP_ESCAPED_HOST
1.853 5.700 0.047 0.99 0.32 SUBJ_REMOVE
0.032 0.098 0.001 0.99 1.00 MIME_BOUND_MA
1.687 0.019 2.470 0.01 -3.13 PGP_SIGNATURE
0.031 0.094 0.001 0.99 1.00 NO_AGE
6.138 0.081 8.983 0.01 -1.00 EMAIL_ATTRIBUTION
0.225 0.693 0.005 0.99 4.53 TO_NO_USER
0.490 1.507 0.013 0.99 1.50 MONEY_BACK
0.744 2.284 0.020 0.99 1.00 MLM
0.053 0.163 0.001 0.99 1.00 HERBAL_VIAGRA
0.718 2.204 0.019 0.99 1.00 BEST_PORN
0.216 0.664 0.005 0.99 1.00 LOSE_POUNDS
0.027 0.083 0.001 0.99 0.55 TRACE_BY_SSN
0.164 0.503 0.004 0.99 4.80 MAIL_IN_ORDER_FORM
0.229 0.704 0.006 0.99 4.29 WANTS_CREDIT_CARD
0.207 0.637 0.005 0.99 1.00 UNSECURED_CREDIT
0.448 1.375 0.013 0.99 1.00 X_LIBRARY
0.255 0.782 0.007 0.99 1.00 FREE_MEMBERSHIP
0.024 0.074 0.001 0.99 3.05 INVESTOR_SPEC_SHEET
0.089 0.272 0.002 0.99 4.20 SUBJ_2_CREDIT
0.208 0.638 0.006 0.99 0.15 EXCUSE_13
4.121 12.600 0.139 0.99 2.00 SPAM_PHRASE_21_34
0.303 0.927 0.009 0.99 3.37 BUGGY_CGI
0.045 0.137 0.001 0.99 1.00 LESBIAN
0.203 0.623 0.006 0.99 4.67 VIAGRA
0.758 2.319 0.024 0.99 1.00 DO_IT_TODAY
0.535 1.639 0.017 0.99 1.00 ADVERT_CODE2
2.604 7.959 0.089 0.99 1.00 REMOVE_SUBJ
5.078 0.084 7.423 0.01 -1.00 X_MAILING_LIST
0.418 1.278 0.013 0.99 1.00 EXTRA_CASH
2.147 6.560 0.075 0.99 1.00 TABLE_THICK_BORDER
8.927 0.156 13.047 0.01 -1.00 USER_AGENT
0.021 0.064 0.001 0.99 1.00 CHILD_SUPPORT
0.637 0.010 0.932 0.01 -4.00 DEBIAN_BTS_BUG
0.355 1.086 0.012 0.99 4.39 FROM_STARTS_WITH_NUMS
0.039 0.120 0.001 0.99 3.00 SPAM_PHRASE_55_XX
3.621 0.064 5.292 0.01 -2.00 SIGNATURE_SHORT_DENSE
0.364 1.112 0.013 0.99 1.00 HAIR_LOSS
0.737 2.249 0.027 0.99 0.48 EXCUSE_15
0.343 1.047 0.012 0.99 2.00 RATWARE_HASH_2
0.384 1.172 0.014 0.99 2.46 ONE_TIME_MAILING
0.054 0.165 0.002 0.99 1.00 NIGERIAN_TRANSACTION_2
1.787 5.444 0.070 0.99 4.43 FRONTPAGE
0.054 0.164 0.002 0.99 2.57 IMPOTENCE
0.070 0.213 0.002 0.99 1.00 WHY_PAY_MORE
0.052 0.159 0.002 0.99 1.00 FORWARD_LOOKING
0.409 1.246 0.016 0.99 1.28 PENIS_ENLARGE
0.205 0.626 0.008 0.99 3.66 DATE_IN_FUTURE_24_48
0.452 1.376 0.018 0.99 0.73 REPLY_REMOVE_SUBJECT
20.830 0.458 30.398 0.01 -1.00 QUOTED_EMAIL_TEXT
0.710 2.159 0.030 0.99 1.51 HTTP_USERNAME_USED
0.048 0.146 0.002 0.99 1.00 AMAZING_STUFF
0.479 1.456 0.020 0.99 1.50 GUARANTEE
0.016 0.050 0.001 0.99 1.00 GET_STARTED_NOW
0.770 2.338 0.033 0.99 1.00 ADULT_SITE
2.235 6.783 0.099 0.99 1.88 NO_OBLIGATION
0.032 0.097 0.001 0.99 2.00 RATWARE_IMKTG
0.509 1.548 0.022 0.99 2.29 ALL_NATURAL
1.624 4.923 0.074 0.99 3.42 COMPLETELY_FREE
0.467 1.416 0.021 0.99 0.85 THIS_AINT_SPAM
0.129 0.391 0.005 0.99 2.89 DATE_IN_FUTURE_48_96
1.541 4.670 0.071 0.98 3.89 MIME_MISSING_BOUNDARY
0.030 0.090 0.001 0.99 2.47 UNIVERSITY_DIPLOMAS
0.043 0.132 0.002 0.99 1.00 BUY_DIRECT
0.399 1.209 0.018 0.99 1.00 VACATION_SCAM
0.068 0.208 0.003 0.99 1.00 AMATEUR_PORN
2.026 6.128 0.100 0.98 1.00 SUBJ_FREE_CAP
5.453 16.463 0.282 0.98 1.08 EXCUSE_3
2.169 6.555 0.110 0.98 2.06 BASE64_ENC_TEXT
0.014 0.043 0.001 0.99 3.06 X_MAILER_GIBBERISH
0.616 0.014 0.898 0.02 -1.00 HOTMAIL_FOOTER2
3.113 9.402 0.160 0.98 1.00 PRIORITY_NO_NAME
0.141 0.427 0.007 0.98 1.00 GUARANTEED_STUFF
1.812 0.044 2.642 0.02 -0.50 SIGNATURE_LONG_SPARSE
1.621 4.895 0.083 0.98 2.60 REMOVE_IN_QUOTES
0.100 0.302 0.005 0.98 2.00 RATWARE_JPFREE
0.062 0.188 0.003 0.98 2.00 RATWARE_MBOMBER
0.073 0.221 0.004 0.98 3.78 STOCK_PICK
0.037 0.112 0.002 0.98 1.00 CENTS_ON_DOLLAR
0.025 0.075 0.001 0.98 2.36 WWW_REMOVEYOU_COM
0.117 0.355 0.006 0.98 1.00 FIND_ANYTHING
1.805 5.435 0.101 0.98 1.00 BAD_CREDIT
0.249 0.752 0.013 0.98 1.75 NUMERIC_HTTP_ADDR
0.227 0.684 0.012 0.98 1.00 LIVE_PORN
0.070 0.212 0.004 0.98 1.00 MIME_BOUND_DIGITS_4
0.123 0.370 0.007 0.98 0.66 X_ESMTP
0.309 0.929 0.018 0.98 2.50 MONEY_MAKING
2.660 7.984 0.160 0.98 1.00 MORTGAGE_OBFU
0.066 0.197 0.004 0.98 1.00 BARELY_LEGAL
1.084 3.253 0.065 0.98 1.00 DIET
1.098 3.294 0.067 0.98 2.29 US_DOLLARS_3
0.464 1.392 0.028 0.98 1.07 OPPORTUNITY
0.128 0.384 0.008 0.98 2.60 REPLY_TO_EMPTY
0.316 0.945 0.020 0.98 2.00 SUBJ_ENDS_IN_SPACE
4.924 14.700 0.332 0.98 1.00 CLICK_BELOW_CAPS
0.020 0.061 0.001 0.98 1.00 HIDDEN_CHARGES
3.310 0.107 4.814 0.02 -1.00 X_LOOP
3.344 9.975 0.229 0.98 2.38 DATE_IN_FUTURE_06_12
2.335 6.968 0.159 0.98 1.00 HTML_FONT_COLOR_GREEN
0.628 1.877 0.042 0.98 4.21 REMOVAL_INSTRUCTIONS
0.113 0.339 0.007 0.98 0.10 MICROSOFT_EXECUTABLE
0.019 0.058 0.001 0.98 1.00 SERIOUS_CASH
1.002 2.991 0.068 0.98 4.30 LIMITED_TIME_ONLY
0.147 0.440 0.010 0.98 1.00 SUBJ_DOLLARS
0.083 0.003 0.121 0.02 -1.00 USER_AGENT_MACOE
0.226 0.676 0.015 0.98 1.00 FREE_ACCESS
0.078 0.003 0.114 0.02 -1.00 FAILURE_NOTICE_2
0.756 2.249 0.055 0.98 2.02 UNSUB_SCRIPT
0.299 0.891 0.021 0.98 1.00 ITS_LEGAL
0.424 1.261 0.030 0.98 3.09 DEAR_FRIEND
0.027 0.080 0.002 0.98 1.00 CONFIDENTIAL_ORDER
0.253 0.753 0.018 0.98 1.00 SAVINGS
0.035 0.104 0.002 0.98 1.00 SEX_FEST
0.035 0.104 0.002 0.98 1.00 SECRET_RECORD
0.810 2.407 0.060 0.98 4.24 NO_EXPERIENCE
0.068 0.204 0.005 0.98 1.00 PRIZE
0.263 0.782 0.019 0.98 1.00 DOMAIN_4U2
0.033 0.099 0.002 0.98 1.00 CASH_BONUS
0.041 0.123 0.003 0.98 1.00 MIME_BOUND_HEX_24
0.176 0.521 0.013 0.98 1.00 UNLIMITED
0.228 0.677 0.018 0.97 2.00 OBFUSCATING_COMMENT
0.273 0.810 0.021 0.97 4.47 US_DOLLARS
0.693 2.051 0.056 0.97 0.88 BULK_EMAIL
0.070 0.209 0.005 0.97 4.43 FORGED_JUNO_RCVD
0.039 0.115 0.003 0.97 1.00 BANKRUPTCY
0.394 1.163 0.032 0.97 1.60 FROM_AND_TO_SAME_3
1.238 3.646 0.107 0.97 1.00 HTML_FONT_COLOR_MAGENTA
0.022 0.066 0.002 0.97 1.80 ONCE_IN_LIFETIME
1.340 3.940 0.118 0.97 1.00 ACT_NOW
0.313 0.922 0.027 0.97 1.00 HIDE_WIN_STATUS
0.402 1.183 0.035 0.97 1.00 INSTANT_ACCESS
0.407 1.197 0.036 0.97 3.00 FOR_INSTANT_ACCESS
0.181 0.533 0.016 0.97 1.00 RICH
0.198 0.583 0.018 0.97 1.99 THE_FOLLOWING_FORM
0.130 0.383 0.012 0.97 2.90 EXCUSE_12
1.549 4.544 0.142 0.97 3.16 WORK_AT_HOME
0.284 0.835 0.025 0.97 4.05 STRONG_BUY
0.155 0.457 0.014 0.97 1.00 ONE_TIME
0.862 2.525 0.081 0.97 1.00 SUB_FREE_OFFER
0.030 0.001 0.043 0.03 0.13 MAJORDOMO
1.757 5.140 0.168 0.97 2.08 OPT_IN
1.095 3.203 0.105 0.97 1.30 EXCUSE_7
0.071 0.208 0.007 0.97 1.83 HTTP_NUMBER_WORD
0.007 0.019 0.001 0.97 1.00 PHONE_CANCER
0.200 0.583 0.019 0.97 1.00 HTML_WIN_OPEN
0.235 0.687 0.023 0.97 4.32 PENIS_ENLARGE2
0.057 0.166 0.005 0.97 2.62 NO_CATCH
0.180 0.525 0.018 0.97 4.41 YOUR_INCOME
7.309 21.265 0.755 0.97 3.31 NORMAL_HTTP_TO_IP
3.399 9.885 0.352 0.97 1.00 MISSING_OUTLOOK_NAME
0.182 0.009 0.264 0.03 -2.40 YAHOO_MSGID_ADDED
0.584 1.701 0.060 0.97 1.00 DATE_MISSING
0.055 0.160 0.005 0.97 1.00 USER_4U2
0.061 0.177 0.006 0.97 3.06 BILLION_DOLLARS
0.018 0.053 0.002 0.97 1.00 LUXURY_CAR
0.973 2.825 0.104 0.96 0.36 EXCUSE_10
0.012 0.035 0.001 0.97 1.00 MIME_BOUND_DIGITS_1
2.542 7.365 0.277 0.96 1.00 HTML_FONT_COLOR_NAME
0.256 0.744 0.027 0.96 1.00 ACCEPT_CREDIT_CARDS
0.177 0.515 0.019 0.96 2.33 FULL_REFUND
0.040 0.117 0.004 0.97 1.00 HTML_90_100
0.040 0.117 0.004 0.97 1.00 RAPE
0.292 0.845 0.032 0.96 2.76 AMAZING
0.011 0.032 0.001 0.96 3.60 EJACULATION
1.598 4.602 0.187 0.96 1.00 HTML_FONT_COLOR_CYAN
0.069 0.200 0.008 0.96 1.47 FROM_NO_USER
0.095 0.273 0.011 0.96 1.00 FREE_CELL_PHONE
0.535 1.540 0.063 0.96 1.00 EXTRA_MPART_TYPE
2.593 7.449 0.312 0.96 1.00 HTML_FONT_COLOR_YELLOW
0.011 0.031 0.001 0.96 1.75 AUTO_EMAIL_REMOVAL
0.752 2.162 0.090 0.96 2.88 HOME_EMPLOYMENT
0.005 0.015 0.001 0.96 1.00 BALANCE_FOR_LONG
0.325 0.934 0.039 0.96 1.00 HTML_FONT_COLOR_UNKNOWN
0.587 1.686 0.071 0.96 1.00 SAVE_UP_TO
6.476 18.536 0.812 0.96 1.50 SPAM_PHRASE_13_21
5.965 17.051 0.758 0.96 1.00 INVALID_DATE
0.361 1.029 0.047 0.96 4.34 EARN_PER_WEEK
0.029 0.083 0.004 0.96 1.00 PROMOTION
0.081 0.231 0.010 0.96 1.00 FREE_SAMPLE
2.127 6.054 0.283 0.96 1.13 FOR_FREE
0.108 0.308 0.014 0.96 1.00 LONG_DISTANCE
1.351 3.844 0.181 0.96 1.94 NO_COST
3.310 9.399 0.450 0.95 1.90 MAILTO_WITH_SUBJ
10.027 28.415 1.391 0.95 0.85 CLICK_HERE_LINK
0.023 0.066 0.003 0.96 2.73 URGENT_BIZ
2.461 6.970 0.343 0.95 0.40 EXCUSE_14
0.018 0.052 0.002 0.96 1.00 FREE_PREVIEW
1.797 5.087 0.251 0.95 1.00 SAVE_MONEY
1.534 4.343 0.214 0.95 0.50 UNDISC_RECIPS
9.999 28.243 1.430 0.95 1.00 HTML_FONT_COLOR_RED
0.075 0.005 0.108 0.05 -1.00 FAILURE_NOTICE_1
0.005 0.013 0.001 0.96 1.00 MONSTERHUT
0.913 2.582 0.129 0.95 1.00 FORGED_RCVD_TRAIL
0.699 1.979 0.099 0.95 1.02 VERY_SUSP_RECIPS
0.040 0.112 0.005 0.95 1.00 SUPPLIES_LIMITED
0.133 0.377 0.019 0.95 1.00 TARGETED
1.524 4.300 0.221 0.95 0.21 SUSPICIOUS_RECIPS
0.085 0.241 0.012 0.95 0.10 MIME_SUSPECT_NAME
13.496 37.952 2.010 0.95 1.67 CTYPE_JUST_HTML
1.241 3.493 0.184 0.95 0.26 EMAIL_MARKETING
0.379 1.065 0.057 0.95 1.00 WHY_WAIT
1.477 4.136 0.228 0.95 1.85 INVALID_MSGID
3.208 8.977 0.499 0.95 0.70 MIME_LONG_LINE_QP
13.128 36.684 2.065 0.95 2.06 BIG_FONT
1.380 3.863 0.213 0.95 1.39 PORN_4
3.926 10.976 0.615 0.95 2.58 FROM_HAS_MIXED_NUMS
1.367 3.818 0.216 0.95 1.00 MANY_EXCLAMATIONS
0.051 0.143 0.008 0.95 1.00 NAME_BRAND
0.078 0.219 0.012 0.95 1.00 GET_IT_NOW
0.054 0.152 0.008 0.95 1.00 SEDUCTION
0.359 1.002 0.057 0.95 3.34 JAVASCRIPT_VERY_UNSAFE
0.039 0.108 0.006 0.95 1.00 NO_FORMS
0.050 0.139 0.008 0.95 1.00 MIME_BOUND_DIGITS_7
0.555 1.544 0.091 0.94 1.00 ALL_CAP_PORN
0.049 0.137 0.008 0.95 0.69 NO_DISSAPOINTMENT
0.212 0.591 0.035 0.94 4.00 RCVD_FAKE_HELO_DOTCOM
0.258 0.717 0.042 0.94 1.00 WHILE_SUPPLIES
0.289 0.801 0.048 0.94 0.50 SUB_HELLO
0.011 0.031 0.002 0.94 1.97 SAFEGUARD_NOTICE
0.029 0.081 0.005 0.94 1.00 FRIEND_AT_PUBLIC
0.392 1.083 0.067 0.94 1.00 CARRIAGE_RETURNS
0.275 0.760 0.047 0.94 0.77 SEE_FOR_YOURSELF
0.249 0.690 0.042 0.94 1.00 OPPORTUNITY_2
0.029 0.080 0.005 0.94 1.00 FREE_WEBSITE
0.998 2.752 0.174 0.94 2.43 MIME_EXCESSIVE_QP
0.028 0.079 0.005 0.94 1.00 HTML_WIN_BLUR
0.011 0.030 0.002 0.94 1.68 EXCUSE_11
3.884 10.684 0.691 0.94 1.00 HTML_FONT_COLOR_UNSAFE
0.035 0.097 0.006 0.94 1.00 OFFSHORE_SCAM
0.735 2.024 0.130 0.94 0.50 MIME_HTML_NO_CHARSET
0.429 1.180 0.076 0.94 1.00 SATISFACTION
2.987 8.203 0.537 0.94 1.00 HTML_FONT_INVISIBLE
0.230 0.633 0.041 0.94 1.00 CONGRATULATIONS
0.348 0.956 0.063 0.94 2.15 MSGID_HAS_NO_AT
2.887 7.895 0.534 0.94 1.00 OFFER
1.394 3.810 0.259 0.94 1.00 HTML_70_90
0.023 0.064 0.004 0.94 1.62 X_SMTPEXP_VERSION
0.023 0.064 0.004 0.94 1.62 X_SMTPEXP_REGISTRATION
0.132 0.361 0.024 0.94 4.46 CHECK_OR_MONEY_ORDER
18.872 51.323 3.630 0.93 1.53 CLICK_BELOW
7.269 19.774 1.396 0.93 1.00 HTML_50_70
0.081 0.222 0.015 0.94 1.00 FANTASTIC
0.061 0.168 0.012 0.94 1.00 LINK_TO_NO_SCHEME
0.528 1.438 0.101 0.93 0.29 X_MSMAIL_PRIORITY_HIGH
0.023 0.062 0.004 0.94 1.00 DRASTIC_REDUCED
3.300 8.958 0.643 0.93 1.94 UPPERCASE_25_50
3.767 10.223 0.735 0.93 1.35 HTML_WITH_BGCOLOR
0.246 0.669 0.047 0.93 0.42 HTML_EMBEDS
3.314 8.977 0.654 0.93 0.25 SUBJ_ALL_CAPS
0.174 0.473 0.034 0.93 1.00 NO_PURCHASE
4.833 0.503 6.867 0.07 1.00 NOSPAM_INC
0.112 0.303 0.022 0.93 2.00 RATWARE_GROUPMAIL
0.099 0.270 0.019 0.93 3.97 MSG_ID_ADDED_BY_MTA
0.101 0.275 0.020 0.93 1.00 WINNER
0.043 0.116 0.008 0.93 1.00 NO_STRINGS
1.364 3.670 0.280 0.93 0.90 X_PRIORITY_HIGH
0.009 0.025 0.002 0.93 1.00 NEW_CUSTOMER
2.542 6.812 0.536 0.93 1.00 ONLY_COST
0.702 1.878 0.149 0.93 1.00 NO_FEE
1.032 2.758 0.221 0.93 1.00 HOT_NASTY
0.014 0.039 0.003 0.93 2.90 MANY_FROMS
0.003 0.008 0.001 0.93 1.00 MORE_TRAFFIC
3.244 8.642 0.709 0.92 1.00 HTML_FONT_COLOR_GRAY
0.730 1.941 0.161 0.92 2.39 CASHCASHCASH
0.930 0.110 1.315 0.08 -1.00 RESENT_TO
0.198 0.023 0.280 0.08 -1.00 USER_AGENT_APPLEMAIL
0.528 1.402 0.118 0.92 1.92 FOR_JUST_SOME_AMT
0.545 1.447 0.122 0.92 1.00 LOW_PRICE
0.887 2.353 0.199 0.92 1.79 ORDER_NOW
0.109 0.289 0.024 0.92 0.95 MASS_EMAIL
0.242 0.637 0.057 0.92 1.00 YOU_WON
0.239 0.626 0.057 0.92 3.29 DIRECT_EMAIL
3.019 7.889 0.732 0.92 0.37 MAILTO_TO_REMOVE
10.900 28.381 2.689 0.91 0.38 FROM_ENDS_IN_NUMS
2.163 5.629 0.536 0.91 0.48 LINES_OF_YELLING_3
0.178 0.464 0.044 0.91 4.20 BE_AMAZED
0.939 2.443 0.233 0.91 0.69 JAVASCRIPT_UNSAFE
0.178 0.462 0.044 0.91 2.99 UPPERCASE_75_100
0.080 0.209 0.020 0.91 1.00 NO_COMBINE
0.002 0.006 0.001 0.91 1.00 MIME_BOUND_EQS_DASHES
0.002 0.006 0.001 0.91 1.00 WEALTH
9.034 23.342 2.315 0.91 1.00 HTML_FONT_COLOR_BLUE
0.151 0.391 0.038 0.91 2.49 DOMAIN_BODY
0.165 0.426 0.043 0.91 1.00 FREE_DVD
1.595 4.098 0.420 0.91 2.33 EXCUSE_1
5.212 13.359 1.386 0.91 1.00 HTML_FONT_FACE_ODD
0.146 0.374 0.039 0.91 3.67 FROM_MALFORMED
0.570 1.459 0.153 0.90 0.94 DATE_IN_PAST_96_XX
2.094 5.341 0.569 0.90 1.64 DATE_IN_PAST_06_12
3.870 9.862 1.055 0.90 0.07 LINES_OF_YELLING_2
0.118 0.302 0.032 0.90 1.81 PROFITS
0.397 1.011 0.108 0.90 1.00 LARGE_COLLECTION
1.640 4.166 0.454 0.90 1.70 DATE_IN_PAST_12_24
0.823 2.087 0.230 0.90 1.00 HTML_FONT_COLOR_NOHASH
0.312 0.048 0.436 0.10 -1.00 APPROVED_BY
0.210 0.032 0.294 0.10 -1.00 HOTMAIL_FOOTER5
0.427 1.081 0.121 0.90 2.00 RATWARE_EGROUPS
0.051 0.129 0.015 0.90 1.00 MIME_BOUND_DIGITS_2
0.135 0.341 0.039 0.90 1.00 GET_PAID
3.765 9.423 1.108 0.89 1.45 FORGED_YAHOO_RCVD
0.214 0.035 0.299 0.10 -1.00 MSN_FOOTER1
0.107 0.268 0.032 0.89 2.41 NO_QS_ASKED
0.012 0.031 0.004 0.89 1.00 LYING_EYES
0.334 0.832 0.100 0.89 2.97 UPPERCASE_50_75
0.065 0.163 0.019 0.89 1.00 GETAWAY
0.002 0.005 0.001 0.89 1.00 CHANGE_TERMS
5.124 12.707 1.562 0.89 0.10 WEB_BUGS
0.227 0.564 0.069 0.89 0.25 ASKS_BILLING_ADDRESS
3.029 7.485 0.936 0.89 0.50 FORGED_HOTMAIL_RCVD
0.108 0.268 0.033 0.89 3.46 DATE_IN_FUTURE_96_XX
0.571 1.407 0.178 0.89 1.45 TO_EMPTY
0.023 0.058 0.007 0.89 1.00 CANCEL
0.520 1.278 0.164 0.89 0.45 GREAT_OFFER
8.753 21.387 2.820 0.88 0.50 SPAM_PHRASE_08_13
0.510 1.246 0.164 0.88 1.40 TONER
1.221 2.961 0.404 0.88 0.76 PLING_PLING
7.751 18.673 2.622 0.88 0.78 MAILTO_LINK
0.070 0.170 0.024 0.88 1.00 WINNING_CAP
0.752 1.810 0.255 0.88 3.42 DATE_IN_FUTURE_03_06
1.916 4.605 0.653 0.88 0.68 DATE_IN_PAST_03_06
8.577 20.524 2.966 0.87 0.01 LINES_OF_YELLING
0.332 0.796 0.114 0.87 1.09 DEAR_SOMETHING
1.371 3.279 0.474 0.87 3.00 TO_HAS_SPACES
0.228 0.045 0.314 0.13 -1.00 HOTMAIL_FOOTER1
0.566 1.354 0.196 0.87 1.00 SAVE_BUCKS
0.264 0.629 0.093 0.87 1.00 GIVING_AWAY
0.044 0.009 0.060 0.13 -1.00 REG_THANKS
0.096 0.227 0.035 0.87 1.00 WINNER_CAP
3.748 8.815 1.368 0.87 1.60 MSG_ID_ADDED_BY_MTA_2
0.022 0.052 0.008 0.87 1.00 NO_INVESTMENT
0.354 0.835 0.129 0.87 1.62 ALL_CAPS_HEADER
0.129 0.303 0.047 0.87 0.40 INCREASE_SOMETHING
0.573 1.344 0.211 0.86 1.00 SALE
0.002 0.004 0.001 0.86 1.00 REMOVE_ES_02
0.631 1.472 0.236 0.86 2.61 DATE_IN_PAST_24_48
0.103 0.240 0.038 0.86 1.00 INVESTMENT
0.186 0.433 0.070 0.86 1.00 NIGERIAN_TRANSACTION_1
1.169 2.709 0.445 0.86 2.65 HTTP_WITH_EMAIL_IN_URL
15.935 3.630 21.715 0.14 -2.50 KNOWN_MAILING_LIST
0.075 0.173 0.028 0.86 1.00 MIME_BOUND_DIGITS_3
0.103 0.239 0.039 0.86 2.50 LARGE_HEX
1.885 4.348 0.729 0.86 2.59 UNSUB_PAGE
0.346 0.797 0.134 0.86 1.00 MONTH_TRIAL
3.988 0.924 5.427 0.15 -1.00 X_ACCEPT_LANG
1.353 3.107 0.529 0.85 1.00 HTML_FONT_FACE_BAD
0.222 0.509 0.087 0.85 1.00 SIGN_UP
0.021 0.049 0.008 0.85 1.00 JOIN_MILLIONS
0.038 0.009 0.052 0.15 -1.00 HOTMAIL_FOOTER4
1.536 3.495 0.615 0.85 0.26 WEIRD_PORT
0.049 0.111 0.019 0.85 1.00 MARKET_SOLUTION
0.056 0.128 0.023 0.85 0.80 DOMAIN_SUBJECT
0.329 0.742 0.134 0.85 1.00 OFFER_EXPIRE
0.847 0.209 1.146 0.15 1.00 SUBJECT_IS_LIST
0.537 1.205 0.223 0.84 1.00 MEMBER_2
0.016 0.036 0.007 0.84 1.00 COUPON
0.204 0.052 0.276 0.16 -1.00 USER_AGENT_ENTOURAGE
0.003 0.006 0.001 0.84 1.80 EU_EMAIL_OPTOUT
0.491 1.089 0.210 0.84 1.00 PLING_QUERY
1.409 3.123 0.605 0.84 1.07 MISSING_HEADERS
2.439 5.333 1.080 0.83 0.69 CALL_FREE
0.880 1.923 0.390 0.83 2.10 SMTPD_IN_RCVD
1.185 2.582 0.529 0.83 0.50 FROM_AND_TO_SAME_1
0.136 0.294 0.062 0.83 1.00 MEMBER
1.262 2.718 0.579 0.82 1.00 TO_ADDRESS_EQ_REAL
21.082 45.096 9.803 0.82 0.50 NO_REAL_NAME
0.059 0.126 0.027 0.82 2.18 EXCUSE_6
0.646 1.365 0.308 0.82 0.03 RELAYING_FRAME
0.229 0.484 0.110 0.82 0.97 GAPPY_SUBJECT
4.213 1.314 5.574 0.19 -0.70 X_AUTH_WARNING
3.791 7.887 1.867 0.81 0.52 MAILTO_TO_SPAM_ADDR
0.166 0.343 0.083 0.81 1.00 MARKETING
1.934 3.965 0.980 0.80 1.52 JAVASCRIPT
0.014 0.030 0.007 0.80 1.00 NO_REFUND
0.644 1.310 0.331 0.80 1.00 HTML_FONT_FACE_CAPS
0.289 0.587 0.149 0.80 1.00 SOCIAL_SEC_NUMBER
0.029 0.059 0.015 0.80 1.00 SHOPPING_SPREE
0.151 0.306 0.079 0.80 2.60 SEARCH_ENGINE_PROMO
0.246 0.490 0.131 0.79 0.10 MSGID_CHARS_WEIRD
0.005 0.009 0.002 0.79 1.00 SIGNIFICANT
0.187 0.369 0.101 0.78 1.00 EXPECT_TO_EARN
0.155 0.055 0.201 0.22 -2.00 FORGOTTEN_PASSWORD
2.378 4.645 1.314 0.78 0.32 EXCUSE_16
1.362 2.631 0.766 0.77 1.11 TO_MALFORMED
4.436 8.537 2.510 0.77 0.21 MSG_ID_ADDED_BY_MTA_3
0.066 0.126 0.037 0.77 2.35 DEAR_EMAIL
1.278 2.456 0.725 0.77 0.50 ASCII_FORM_ENTRY
1.350 2.591 0.767 0.77 1.00 FREE_QUOTE
0.068 0.128 0.039 0.76 2.60 COMMUNIGATE
1.830 0.742 2.341 0.24 -0.50 SUBJECT_MONTH
0.002 0.004 0.001 0.76 1.00 CLICK_TO_REMOVE_3
0.378 0.704 0.224 0.76 1.00 US_DOLLARS_4
1.807 3.360 1.077 0.76 1.93 RISK_FREE
1.611 2.989 0.964 0.76 0.01 FORGED_RCVD_FOUND
0.140 0.259 0.084 0.76 1.00 RESERVES_RIGHT
0.195 0.360 0.118 0.75 2.06 MSGID_CHARS_SPAM
0.007 0.013 0.004 0.75 1.00 SLASH_PRICE
0.424 0.769 0.262 0.75 0.50 FROM_AND_TO_SAME_5
2.387 1.046 3.017 0.26 -1.00 USER_AGENT_MOZILLA_XM
0.620 1.108 0.391 0.74 1.00 FREE_TRIAL
1.069 1.905 0.677 0.74 0.18 US_DOLLARS_2
0.012 0.021 0.007 0.74 1.00 FREE_HOSTING
0.017 0.008 0.022 0.26 -5.00 LISTBUILDER
0.133 0.236 0.084 0.74 0.10 JAVASCRIPT_URI
4.095 1.891 5.130 0.27 -2.00 EXCHANGE_SERVER
3.472 5.834 2.363 0.71 0.49 DEAR_SOMEBODY
2.080 1.060 2.559 0.29 -0.50 SUBJECT_MONTH_2
0.392 0.646 0.272 0.70 1.40 FROM_AND_TO_SAME_6
1.287 2.091 0.910 0.70 1.00 HTML_COMMENT_UNIQUE_ID
6.889 11.118 4.903 0.69 0.38 SUPERLONG_LINE
7.412 11.859 5.324 0.69 0.10 SPAM_PHRASE_05_08
0.978 0.540 1.184 0.31 -1.00 FWD_MSG
1.182 1.870 0.858 0.69 1.00 DISCLAIMER
2.782 4.309 2.065 0.68 1.00 MAY_BE_FORGED
0.033 0.050 0.024 0.67 3.09 DATE_WARNING
0.044 0.067 0.033 0.67 1.56 DONT_DELETE
0.220 0.334 0.166 0.67 2.48 DATE_IN_PAST_48_96
4.705 2.820 5.591 0.34 -0.50 USER_AGENT_OUTLOOK
0.918 0.549 1.092 0.33 -1.00 SUBJECT_FREQ
0.168 0.250 0.129 0.66 1.00 GIFT_CERTIFICATE
0.218 0.317 0.172 0.65 -2.50 FROM_AND_TO_SAME_2
0.500 0.725 0.394 0.65 0.40 SUBJ_MISSING
1.582 2.226 1.279 0.64 0.39 GAPPY_TEXT
1.056 0.735 1.206 0.38 -0.10 SUBJECT_IS_NEWS
0.004 0.003 0.004 0.38 -1.00 TRACK_NUMBER
0.371 0.262 0.422 0.38 -1.00 ORDER_STATUS
0.049 0.064 0.042 0.61 1.00 BIG_BUCKS
0.048 0.062 0.042 0.60 1.00 CANT_LIVE_WITHOUT
2.972 3.791 2.587 0.59 1.00 FREE_MONEY
0.491 0.382 0.542 0.41 -1.00 SUBJECT_HAS_DATE
0.093 0.110 0.085 0.56 1.76 PLEASE_READ
1.600 1.556 1.620 0.49 -0.30 OUTLOOK_FW_MSG
0.116 0.117 0.115 0.50 1.00 MIME_BOUND_DIGITS_5
0.154 0.153 0.154 0.50 -1.00 BALANCE_FOR_LONG_40K
0.124 0.123 0.125 0.50 0.50 COMMENT
0.005 0.005 0.005 0.49 1.00 PRICES_WONT_LAST
0.214 0.224 0.208 0.52 -1.00 ACCOUNT_CLICK
0.029 0.027 0.030 0.48 1.00 FREE_TICKETS
0.080 0.075 0.082 0.48 3.70 EXCUSE_17
0.261 0.230 0.276 0.45 1.00 T_FREE_WEBSITE
9.494 8.256 10.075 0.45 0.10 SPAM_PHRASE_03_05
0.156 0.188 0.141 0.57 -1.00 MAILER_DAEMON
0.741 0.589 0.812 0.42 1.71 SPAM_REDIRECTOR
0.002 0.001 0.002 0.42 2.03 URI_IS_POUND
0.026 0.018 0.030 0.38 5.00 FORGED_EBAY_RCVD
0.757 1.047 0.620 0.63 -1.00 BALANCE_FOR_LONG_20K
0.508 0.297 0.607 0.33 4.26 FROM_MISSING
1.221 1.888 0.907 0.68 -0.10 USER_AGENT_AOL
0.126 0.072 0.151 0.32 1.00 MIME_BOUND_DIGITS_6
9.083 14.646 6.470 0.69 -1.00 USER_AGENT_OE
0.012 0.006 0.015 0.30 0.50 OUTLOOK_UNDISC_RECIPS
0.896 0.420 1.119 0.27 -0.97 MIME_NULL_BLOCK
8.286 3.760 10.412 0.27 -0.05 SPAM_PHRASE_02_03
41.099 18.305 51.805 0.26 -0.20 SPAM_PHRASE_00_01
6.987 3.068 8.828 0.26 1.00 TO_BE_REMOVED_REPLY
0.064 0.120 0.038 0.76 -1.00 PRIVACY_STATEMENT
0.048 0.090 0.028 0.76 -1.00 T_NASD_FINANCIAL
1.843 0.622 2.417 0.20 1.00 SUSPECT_LIST_HEADERS
2.074 0.685 2.726 0.20 0.48 TO_LOCALPART_EQ_REAL
11.879 3.688 15.725 0.19 -0.10 SPAM_PHRASE_01_02
0.672 1.468 0.299 0.83 -1.00 USER_AGENT_THEBAT
0.002 0.004 0.001 0.86 -1.00 MAILBITS_EMAIL
0.071 0.012 0.099 0.10 -2.50 FROM_AND_TO_SAME_4
0.003 0.008 0.001 0.93 -0.50 FROM_US_PHONE
4.040 0.034 5.922 0.01 -100.00 USER_IN_WHITELIST
0.000 0.000 0.000 0.00 1.00 REMOVE_ES_06
0.000 0.000 0.000 0.00 1.00 RATWARE_02
0.002 0.000 0.003 0.00 1.00 RATWARE_05
0.000 0.000 0.000 0.00 1.00 SPY_ON_FRIENDS
0.000 0.000 0.000 0.00 1.00 REMOVE_ES_05
0.000 0.000 0.000 0.00 1.00 RATWARE_07
0.000 0.000 0.000 0.00 1.00 REMOVE_ES_04
0.000 0.000 0.000 0.00 2.00 UNDESIRED_LANGUAGE_BODY
0.000 0.000 0.000 0.00 1.00 LIFE_INSURANCE
0.000 0.000 0.000 0.00 -20.00 USER_IN_MORE_SPAM_TO
0.000 0.000 0.001 0.00 1.00 CLICK_TO_REMOVE_MAILTO
0.000 0.000 0.001 0.00 1.00 FREE_FLAG
0.000 0.000 0.000 0.00 4.00 STAINLESS_STEEL
0.000 0.000 0.001 0.00 1.00 SUBSCRIBE_ES_01
0.000 0.000 0.000 0.00 2.00 RATWARE_UPROAR
0.000 0.000 0.000 0.00 1.00 RATWARE_10
0.000 0.000 0.000 0.00 1.00 BUGGY_CGI_ES
0.000 0.000 0.000 0.00 1.00 RATWARE_12
0.000 0.000 0.000 0.00 1.00 RATWARE_13
0.000 0.000 0.000 0.00 1.00 RATWARE_14
0.000 0.000 0.000 0.00 1.00 RATWARE_15
0.000 0.000 0.000 0.00 1.00 RATWARE_30
0.000 0.000 0.000 0.00 1.00 RATWARE_16
0.000 0.000 0.000 0.00 -100.00 USER_IN_ALL_SPAM_TO
0.000 0.000 0.000 0.00 1.00 RATWARE_31
0.000 0.000 0.000 0.00 1.00 RATWARE_18
0.000 0.000 0.000 0.00 1.00 CLICKSFORMONEY_NET
0.000 0.000 0.000 0.00 1.00 RATWARE_19
0.000 0.000 0.000 0.00 1.00 VIGORA
0.000 0.000 0.000 0.00 1.00 BREAKTHROUGH
0.000 0.000 0.000 0.00 1.00 RATWARE_32
0.000 0.000 0.000 0.00 1.00 RATWARE_33
0.000 0.000 0.000 0.00 1.00 RATWARE_34
0.000 0.000 0.000 0.00 1.00 RATWARE_35
0.000 0.000 0.000 0.00 1.00 RATWARE_36
0.000 0.000 0.000 0.00 1.00 BACKED_BY
0.000 0.000 0.000 0.00 1.00 NO_SPENDING
0.000 0.000 0.000 0.00 1.00 RATWARE_37
0.000 0.000 0.000 0.00 1.00 RATWARE_38
0.000 0.000 0.000 0.00 1.00 FROM_UGETMORE
0.001 0.000 0.001 0.00 1.00 BONUS_PAYMENT
0.000 0.000 0.000 0.00 4.00 BUGGY_CGI_DE
0.000 0.000 0.000 0.00 1.00 BRAND_NEW_PAGER
0.000 0.000 0.000 0.00 1.00 WWW_TRAFFICWOW_NET
0.001 0.000 0.001 0.00 1.00 ITS_EFFECTIVE
0.000 0.000 0.000 0.00 3.90 LASER_PRINTER
0.000 0.000 0.000 0.00 1.00 RATWARE_41
0.000 0.000 0.000 0.00 1.00 RATWARE_43
0.000 0.000 0.000 0.00 1.00 RATWARE_20
0.000 0.000 0.000 0.00 1.00 RATWARE_45
0.002 0.000 0.002 0.00 1.00 RATWARE_21
0.017 0.054 0.000 1.00 -1.00 USER_AGENT_GNUS_XM
0.000 0.000 0.000 0.00 3.46 FAKED_UNDISC_RECIPS_AT
0.000 0.000 0.000 0.00 100.00 USER_IN_BLACKLIST
0.000 0.000 0.000 0.00 2.26 CHARSET_FARAWAY_BODY
0.000 0.000 0.000 0.00 1.00 RATWARE_24
0.000 0.000 0.000 0.00 1.00 RATWARE_25
0.000 0.000 0.000 0.00 1.00 RATWARE_26
0.000 0.000 0.000 0.00 1.00 RATWARE_27
0.000 0.000 0.000 0.00 4.00 BUGGY_CGI_PT
0.000 0.000 0.000 0.00 1.00 PRINT_OUT_AND_FAX
0.000 0.000 0.000 0.00 -6.00 USER_IN_WHITELIST_TO
0.000 0.000 0.000 0.00 1.00 REMOVE_ES_07
0.000 0.000 0.000 0.00 1.00 RATWARE_00
0.001 0.000 0.001 0.00 1.00 SHOES_GUY
0.000 0.000 0.000 0.00 1.00 EXCUSE_ES_03
0.000 0.000 0.001 0.00 1.00 EXCUSE_ES_02
0.000 0.000 0.000 0.00 1.00 SELECTED
0.000 0.000 0.000 0.00 1.00 IRS
0.000 0.000 0.001 0.00 0.57 CASINO
0.001 0.000 0.001 0.00 1.00 SLIGHTLY_UNSAFE_JAVASCRIPT
0.022 0.000 0.032 0.00 1.00 DIFFERENT_REPLY_TO
0.001 0.000 0.002 0.00 1.00 FROM_NAME_EQ_FROM_ADDR
0.000 0.000 0.000 0.00 1.00 RATWARE_01
0.067 0.000 0.099 0.00 1.00 SUBJ_HAS_Q_MARK
0.047 0.000 0.068 0.00 1.00 FROM_AND_TO_SAME
0.000 0.000 0.001 0.00 1.00 ALL_CAPS_SUBJECT
0.024 0.000 0.035 0.00 1.00 PORN_10
0.002 0.000 0.003 0.00 1.00 PORN_11
0.002 0.000 0.003 0.00 1.00 PORN_12
0.000 0.000 0.001 0.00 1.00 REAL_THING
0.000 0.000 0.001 0.00 1.00 INCREASE_TRAFFIC
0.000 0.000 0.001 0.00 1.00 VERY_SUSP_CC_RECIPS
0.059 0.000 0.087 0.00 1.00 COPYRIGHT_CLAIMED
0.002 0.000 0.003 0.00 1.00 PORN_3
0.001 0.000 0.002 0.00 1.00 PORN_8
0.017 0.000 0.025 0.00 1.00 SIGNATURE_DELIM
0.000 0.000 0.001 0.00 1.00 INCREASE_SALES
0.002 0.000 0.003 0.00 1.00 UNIFIED_PATCH
0.003 0.000 0.005 0.00 1.00 DATE_IN_FUTURE
0.002 0.000 0.004 0.00 1.00 ONE_HUNDRED_PC_FREE
0.001 0.000 0.001 0.00 1.00 RATWARE
0.000 0.000 0.001 0.00 1.00 INVALID_DATE_ODD_MONTH
0.111 0.000 0.164 0.00 1.00 PLING
0.000 0.000 0.001 0.00 1.00 SUSPICIOUS_CC_RECIPS
0.052 0.000 0.076 0.00 1.00 TO_UNSUB_REPLY
0.010 0.000 0.015 0.00 1.00 X_EM_VER_PRESENT
0.061 0.000 0.090 0.00 1.00 SUBJ_ENDS_IN_Q_MARK
0.000 0.000 0.001 0.00 1.00 mass-check


-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
rODbegbie
2002-08-23 15:08:55 UTC
Permalink
Post by Justin Mason
0.211 0.660 0.000 1.00 1.00 USERNAME_IN_SUBJECT
0.211 0.660 0.000 1.00 1.50 USERNAME_IN_SUBJECT_4
0.211 0.660 0.000 1.00 1.00 USERNAME_IN_SUBJECT_5
0.211 0.660 0.000 1.00 0.50 USERNAME_IN_SUBJECT_6
0.211 0.660 0.000 1.00 0.50 USERNAME_IN_SUBJECT_2
0.211 0.660 0.000 1.00 2.00 USERNAME_IN_SUBJECT_3
Was there *any* difference in the hitrates for these? Can we limit them to
just one now?
Post by Justin Mason
0.017 0.000 0.025 0.00 -20.00 HABEAS_SWE
People are already using the haiku?!
Post by Justin Mason
1.600 1.556 1.620 0.49 -0.30 OUTLOOK_FW_MSG
0.012 0.006 0.015 0.30 0.50 OUTLOOK_UNDISC_RECIPS
Perhaps these can be improved by making them meta rules that require
USER_AGENT_OUTLOOK
Post by Justin Mason
0.000 0.000 0.001 0.00 1.00 mass-check
That rule should *definitely* go ;)

rOD.

--
"The man on the Mac in Macedonia
hits on the girl with the fully-loaded PC in D.C."

Doing the blogging thang again at http://www.groovymother.com/ <<



-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Theo Van Dinter
2002-08-23 15:37:29 UTC
Permalink
Post by rODbegbie
Post by Justin Mason
0.211 0.660 0.000 1.00 1.00 USERNAME_IN_SUBJECT
0.211 0.660 0.000 1.00 1.50 USERNAME_IN_SUBJECT_4
0.211 0.660 0.000 1.00 1.00 USERNAME_IN_SUBJECT_5
0.211 0.660 0.000 1.00 0.50 USERNAME_IN_SUBJECT_6
0.211 0.660 0.000 1.00 0.50 USERNAME_IN_SUBJECT_2
0.211 0.660 0.000 1.00 2.00 USERNAME_IN_SUBJECT_3
Was there *any* difference in the hitrates for these? Can we limit them to
just one now?
did any of the people who were worried about false positives mass-check
or send in examples? given current spam, I'd like 3 to be used.
it's the most strict, so should cause the least false positives.
I'm completely amazed that 5 and 6 didn't get FPs, they were there for
testing and should be disabled.

3 is the most strict, and therefore is the easiest for spammers to get
around (add a space after the comma, change case, etc.) 1 removes the
space option, and 4 removes the case option, while 2 removes both.

so I'd probably go with 3 for right now and see if the spammers change
their messages. surprisingly few spams come in that trigger this rule,
so I'm not too too worried about it. I just don't want to go sproinging
FPs up on people.
Post by rODbegbie
Post by Justin Mason
0.017 0.000 0.025 0.00 -20.00 HABEAS_SWE
People are already using the haiku?!
It's in every mail from Dan Kohn, surprise! Is someone mass checking
the SA list?
--
Randomly Generated Tagline:
McBorg: over 50 billion assimlated
Craig R.Hughes
2002-08-23 21:26:57 UTC
Permalink
I have a lot of non-list mail from Dan

C
Post by Theo Van Dinter
Post by rODbegbie
Post by Justin Mason
0.017 0.000 0.025 0.00 -20.00 HABEAS_SWE
People are already using the haiku?!
It's in every mail from Dan Kohn, surprise! Is someone mass checking
the SA list?
-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Daniel Quinlan
2002-08-26 01:29:24 UTC
Permalink
Post by Theo Van Dinter
Post by rODbegbie
Post by Justin Mason
0.017 0.000 0.025 0.00 -20.00 HABEAS_SWE
People are already using the haiku?!
It's in every mail from Dan Kohn, surprise! Is someone mass checking
the SA list?
It is reckless to assign this -20 and even worse to place it outside
the GA-section. The mark is 100% forgeable and spammers outside the
legal reach of Habeas can use it with impunity to get their spam
through filters.

Also, why does NMS_CGI_NOT_BUGGY have such a low negative score? Does
it trip that much stuff? Why can't this be GA-scored as well?

Dan


-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Bart Schaefer
2002-08-26 02:47:12 UTC
Permalink
Post by Daniel Quinlan
Post by Theo Van Dinter
Post by rODbegbie
Post by Justin Mason
0.017 0.000 0.025 0.00 -20.00 HABEAS_SWE
People are already using the haiku?!
It's in every mail from Dan Kohn, surprise! Is someone mass checking
the SA list?
It is reckless to assign this -20 and even worse to place it outside
the GA-section.
You're missing the HABEAS_HIL rbl check. The idea is that if the mark has
been forged, the forger's IP will be in hil.habeas.com, and consequently
the HABEAS_SWE rule is canceled out; so only non-forged uses of the mark
will get the -20 (and forged uses are treated as definitely spam, as
HABEAS_HIL is +30).

Of course there's inevitably going to be some delay in IPs getting added
to the HIL, but the hope is that if you see forged marks you'll report
them right away, and then any more from that source will get tagged.

HABEAS_SWE has to be big enough to let a pretty spammy-looking message get
by, and HABEAS_HIL has to be at least big enough to cancel HABEAS_SWE.
I'm not sure -20/+30 is right (-15/+20 might be better), but I don't see
how the GA could be used to compute those values.



-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Daniel Quinlan
2002-08-26 06:26:10 UTC
Permalink
Post by Bart Schaefer
You're missing the HABEAS_HIL rbl check. The idea is that if the mark has
been forged, the forger's IP will be in hil.habeas.com, and consequently
the HABEAS_SWE rule is canceled out; so only non-forged uses of the mark
will get the -20 (and forged uses are treated as definitely spam, as
HABEAS_HIL is +30).
Actually, I saw the HABEAS_HIL rbl check.
Post by Bart Schaefer
Of course there's inevitably going to be some delay in IPs getting added
to the HIL, but the hope is that if you see forged marks you'll report
them right away, and then any more from that source will get tagged.
Exactly, except that I think you may be too optimistic about the
reporting, the timeframe, and the accuracy.
Post by Bart Schaefer
HABEAS_SWE has to be big enough to let a pretty spammy-looking message get
by, and HABEAS_HIL has to be at least big enough to cancel HABEAS_SWE.
I'm not sure -20/+30 is right (-15/+20 might be better), but I don't see
how the GA could be used to compute those values.
You can write your own ticket with -20 or -15. Even -10 might be too
much for a single substring that is just asking to be forged. I mean,
just get a dynamic or throwaway IP, send some spam with the mark, then
switch IPs.

Really, we need to find a way to start scoring realtime network tests.
Maybe not the same GA run as everything else, but some adaptive number
based on recent realtime performance.

Dan


-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Sidney Markowitz
2002-08-26 15:48:31 UTC
Permalink
Post by Daniel Quinlan
Exactly, except that I think you may be too optimistic
about the reporting, the timeframe, and the accuracy.
That's another argument for a "HLL" (Habeas Licensee List) which would be less
inclusive than a test for the mark of for HIL, but more timely and accurate. When an
ip address is on the HLL it could get a huge negative score, bigger than the Habeas
mark alone.

-- sidney




-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Craig R.Hughes
2002-08-26 15:31:03 UTC
Permalink
The GA run on the current corpus is yielding an average (mean)
score for false-positives of about 10 points, which is higher
than before. There's some tweaking to be done yet (rule
elimination and such), so that will likely fall, but -15 might
not be enough of a correction in many situations. I will set
the scores for the Habeas stuff and the Ironport stuff to be
sensible given:

a) estimated likeliness of compliance with Habeas/Bonded sender
rules, and
b) score correction level needed to achieve purposes of those projects.

C
Post by Bart Schaefer
Post by Daniel Quinlan
Post by Theo Van Dinter
Post by rODbegbie
Post by Justin Mason
0.017 0.000 0.025 0.00 -20.00 HABEAS_SWE
People are already using the haiku?!
It's in every mail from Dan Kohn, surprise! Is someone mass checking
the SA list?
It is reckless to assign this -20 and even worse to place it outside
the GA-section.
You're missing the HABEAS_HIL rbl check. The idea is that if
the mark has
been forged, the forger's IP will be in hil.habeas.com, and
consequently
the HABEAS_SWE rule is canceled out; so only non-forged uses of the mark
will get the -20 (and forged uses are treated as definitely spam, as
HABEAS_HIL is +30).
Of course there's inevitably going to be some delay in IPs
getting added
to the HIL, but the hope is that if you see forged marks you'll report
them right away, and then any more from that source will get tagged.
HABEAS_SWE has to be big enough to let a pretty spammy-looking
message get
by, and HABEAS_HIL has to be at least big enough to cancel HABEAS_SWE.
I'm not sure -20/+30 is right (-15/+20 might be better), but I
don't see
how the GA could be used to compute those values.
-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
_______________________________________________
Spamassassin-talk mailing list
https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Matt Sergeant
2002-08-27 08:33:02 UTC
Permalink
Post by Daniel Quinlan
Post by Theo Van Dinter
Post by rODbegbie
Post by Justin Mason
0.017 0.000 0.025 0.00 -20.00 HABEAS_SWE
People are already using the haiku?!
It's in every mail from Dan Kohn, surprise! Is someone mass checking
the SA list?
It is reckless to assign this -20 and even worse to place it outside
the GA-section. The mark is 100% forgeable and spammers outside the
legal reach of Habeas can use it with impunity to get their spam
through filters.
Also, why does NMS_CGI_NOT_BUGGY have such a low negative score? Does
it trip that much stuff? Why can't this be GA-scored as well?
The problem with NMS_CGI is that it has to both counteract any score for
BUGGY_CGI (because the body text for NMS formmail is the same as for
Matt's Scripts formmail), and "white" the email. So it has to do two
jobs in one.

It can't be sanely GA scored yet since nobody yet has any corpus data
for it. I had some, but it got lost in a mass deletion.

Matt.



-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Craig R.Hughes
2002-08-23 21:19:32 UTC
Permalink
Those are probably all message from Dan Kohn to me, or to the list.
Post by rODbegbie
Post by Justin Mason
0.017 0.000 0.025 0.00 -20.00 HABEAS_SWE
People are already using the haiku?!
-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Bart Schaefer
2002-08-23 16:20:14 UTC
Permalink
The total corpora size was 166978 non-spam messages, and 77531 spam mails
Curiosity (feel free to point me at a document somewhere):

Has any effort been made to de-dup the 77531 spams? That is, it's not
likely that any given non-spam is in many people's corpora, but spam tends
to be everywhere, and could easily be anywhere more than once as well.
Will matching the same content multiple times increase a rule's score, or
does the algorithm take that into account (e.g., some sort of checksum)?
69 of them hit nothing in any of the corpora, which probably means they're
broken. I recommend
(a) moving them to 70_broken_rules.cf in CVS
(b) commenting those rules in the release and CVS
(c) log a bug against each one in Bugzilla, so people can
track progress on fixing (if they *are* broken) or dropping
them (if they're just obsolete)
Another 43 have a very high FP ratio, matching S/O ratios between 0.0
(that means the rule hits 100% incorrectly) to 0.66 (gives an FP for 1 in
every 3 mails). I suggest we comment these as well, using the protocol
above: alternatively, we could suggest score ranges for these rules, and
then hand-check the scores the GA comes up with to ensure the scores do
not exceed the suggested ranges.
6.889 11.118 4.903 0.69 0.38 SUPERLONG_LINE
7.412 11.859 5.324 0.69 0.10 SPAM_PHRASE_05_08
2.782 4.309 2.065 0.68 1.00 MAY_BE_FORGED
These three give me that itchy feeling, in spite of scoring above 0.66.
0.978 0.540 1.184 0.31 -1.00 FWD_MSG
I'm a bit surprised that any spams hit this one. Did somebody forward a
spam that then went into a corpus?
0.044 0.067 0.033 0.67 1.56 DONT_DELETE
0.220 0.334 0.166 0.67 2.48 DATE_IN_PAST_48_96
I presume those scores would go down on a re-run of the GA ...

So here we get to the ones Jason M. is worried about; I'll mention only
1.582 2.226 1.279 0.64 0.39 GAPPY_TEXT
Gappy text in a non-spam? It must be a H E A D L I N E or something.
Maybe just remove space from the list of "gap" characters?

Or maybe this and the SUSPICIOUS RECIPS rules have become ineffective
because previously they were so effective that spammers have stopped the
practice.
1.600 1.556 1.620 0.49 -0.30 OUTLOOK_FW_MSG
Same Q as above about FWD_MSG ...
0.154 0.153 0.154 0.50 -1.00 BALANCE_FOR_LONG_40K
Compensation rule.
4.040 0.034 5.922 0.01 -100.00 USER_IN_WHITELIST
0.000 0.000 0.000 0.00 -100.00 USER_IN_ALL_SPAM_TO
0.000 0.000 0.000 0.00 100.00 USER_IN_BLACKLIST
0.000 0.000 0.000 0.00 -6.00 USER_IN_WHITELIST_TO
Ahem.
0.000 0.000 0.000 0.00 1.00 REMOVE_ES_06
0.000 0.000 0.000 0.00 1.00 REMOVE_ES_05
0.000 0.000 0.000 0.00 1.00 REMOVE_ES_04
0.000 0.000 0.001 0.00 1.00 SUBSCRIBE_ES_01
0.000 0.000 0.000 0.00 1.00 REMOVE_ES_07
0.000 0.000 0.000 0.00 1.00 EXCUSE_ES_03
0.000 0.000 0.001 0.00 1.00 EXCUSE_ES_02
Anybody got a Spanish corpus?
0.000 0.000 0.000 0.00 2.00 UNDESIRED_LANGUAGE_BODY
I know I get hits on that one. Does mass-check use TextCat?



-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Craig R.Hughes
2002-08-23 21:29:44 UTC
Permalink
You could actually argue this both ways. You might *not* want
to de-dupe, since you want to penalize the more-widely
distributed spam more than narrowly-distributed spam. I think
for simplicity's sake, it's probably best to not do too much
work on de-duping, but people should make sure they're only
scanning their *own* mail once (Justin?) -- be careful when you
do a mass-check >> file.log that you don't still have your old
output in the log.

C
Post by Bart Schaefer
Has any effort been made to de-dup the 77531 spams? That is, it's not
likely that any given non-spam is in many people's corpora, but
spam tends
to be everywhere, and could easily be anywhere more than once as well.
Will matching the same content multiple times increase a rule's
score, or
does the algorithm take that into account (e.g., some sort of
checksum)?
-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Craig R.Hughes
2002-08-23 21:32:57 UTC
Permalink
No it doesn't -- that score is set manually, and I think rightly so.

C
Post by Bart Schaefer
Post by Justin Mason
0.000 0.000 0.000 0.00 2.00 UNDESIRED_LANGUAGE_BODY
I know I get hits on that one. Does mass-check use TextCat?
-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Daniel Quinlan
2002-08-26 01:42:36 UTC
Permalink
Post by Bart Schaefer
Has any effort been made to de-dup the 77531 spams? That is, it's not
likely that any given non-spam is in many people's corpora, but spam tends
to be everywhere, and could easily be anywhere more than once as well.
Will matching the same content multiple times increase a rule's score, or
does the algorithm take that into account (e.g., some sort of checksum)?
In addition to some of the other reasons people have mentioned for not
removing dups, there's also that it would not work for a large amount
of mass-check data: mass-check on folders where each message is a
separate file does not use the message-id, spammers use duplicate
message-ids and sometimes even send the same spam at different times
and dates (so it may look like a dup, but isn't), etc.
Post by Bart Schaefer
Post by Justin Mason
0.978 0.540 1.184 0.31 -1.00 FWD_MSG
I'm a bit surprised that any spams hit this one. Did somebody forward a
spam that then went into a corpus?
My corpus is quite clean and I get spam hits.

0.200 0.200 0.000 1.00 -1.00 FWD_MSG
Post by Bart Schaefer
Post by Justin Mason
0.044 0.067 0.033 0.67 1.56 DONT_DELETE
0.220 0.334 0.166 0.67 2.48 DATE_IN_PAST_48_96
I presume those scores would go down on a re-run of the GA ...
Yes.
Post by Bart Schaefer
Post by Justin Mason
1.582 2.226 1.279 0.64 0.39 GAPPY_TEXT
Gappy text in a non-spam? It must be a H E A D L I N E or something.
Maybe just remove space from the list of "gap" characters?
For 2.50, we could try to figure out which nonspams are hitting and
improve the rule.
Post by Bart Schaefer
Or maybe this and the SUSPICIOUS RECIPS rules have become ineffective
because previously they were so effective that spammers have stopped the
practice.
Same here.
Post by Bart Schaefer
Post by Justin Mason
1.600 1.556 1.620 0.49 -0.30 OUTLOOK_FW_MSG
Same Q as above about FWD_MSG ...
Same answer.

0.500 0.500 0.000 1.00 -0.30 OUTLOOK_FW_MSG
Post by Bart Schaefer
Post by Justin Mason
4.040 0.034 5.922 0.01 -100.00 USER_IN_WHITELIST
This proves my point that we need (needed?) to lower the score for default
whitelist entries. :-)
Post by Bart Schaefer
Post by Justin Mason
0.000 0.000 0.000 0.00 2.00 UNDESIRED_LANGUAGE_BODY
I know I get hits on that one. Does mass-check use TextCat?
Nope, it's a user preference to set the language (and turn it on).
You can't do the test without that preference setting.

Dan


-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Dan Kohn
2002-08-26 06:08:57 UTC
Permalink
As Bart indicated, the HIL is designed to stop people outside of Habeas'
(immediate) reach:
http://www.habeas.com/faq/#5.4

- dan
--
Dan Kohn <mailto:***@dankohn.com>
<http://www.dankohn.com/> <tel:+1-650-327-2600>


-----Original Message-----
From: Daniel Quinlan [mailto:***@pathname.com]
Sent: Sunday, August 25, 2002 6:29 PM
To: spamassassin-***@lists.sourceforge.net
Cc: ***@deersoft.com; rODbegbie; Theo Van Dinter
Subject: Re: [SAtalk] 2.40 RELEASE: oops -- newer freqs.
Post by Theo Van Dinter
Post by rODbegbie
Post by Justin Mason
0.017 0.000 0.025 0.00 -20.00 HABEAS_SWE
People are already using the haiku?!
It's in every mail from Dan Kohn, surprise! Is someone mass checking
the SA list?
It is reckless to assign this -20 and even worse to place it outside the
GA-section. The mark is 100% forgeable and spammers outside the legal
reach of Habeas can use it with impunity to get their spam through
filters.

Also, why does NMS_CGI_NOT_BUGGY have such a low negative score? Does
it trip that much stuff? Why can't this be GA-scored as well?

Dan


-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old cell
phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
_______________________________________________
Spamassassin-talk mailing list Spamassassin-***@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/spamassassin-talk



-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Daniel Quinlan
2002-08-26 19:09:40 UTC
Permalink
Post by Dan Kohn
As Bart indicated, the HIL is designed to stop people outside of
Habeas' (immediate) reach: http://www.habeas.com/faq/#5.4
Yeah, I read the FAQ. It's not exactly a secret that you can change
your IP address after sending out spam. Funny, their FAQ mentions
nothing about IP addresses.

I don't know why so many people are buying into the Habeas marketing
material and assigning huge scores to the mark before we see whether
it will actually work. I think the HIL is a better bet, but only
because we've seen that DNSBLs work well enough to use.

I think it's an interesting idea and I hope it works to reduce spam,
but there is no reason to project hope into large SA scores. We might
as well dump the GA if that's how we're going to operate.

Dan


-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Malte S. Stretz
2002-08-26 19:46:06 UTC
Permalink
Post by Daniel Quinlan
[...]
I don't know why so many people are buying into the Habeas marketing
material and assigning huge scores to the mark before we see whether
it will actually work. [...]
Because it won't work if nobody starts to use it ;-)
Post by Daniel Quinlan
I think it's an interesting idea and I hope it works to reduce spam,
but there is no reason to project hope into large SA scores. [...]
As it's free to use the tag as an individual, we (the SA developers & users)
might support it by adding the tag to our mail. I don't like the clumsy
headers in my mail (would be nice if it was possible to merge line 5+ into
one) but let's have a try.

There comes one question in my mind: The use is free for individuals and
ISPs but not for corporates. I have one account for both my private mail
and the one going to my customers (very small business :). Do I
theoretically have to pay the 200 bucks per year? What is about eg. Mat,
sending "private" (OSS) mail from his MessageLabs address?

Malte

P.S.: Maybe we should create a spamassassin-talk-habeas list ;-)
--
-- Coding is art.
--
-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Craig R.Hughes
2002-08-26 21:12:15 UTC
Permalink
Stupid OSX Mail.app won't let me insert custom headers. There
are exactly 3 features missing from this program which would
make it the best mail program I've ever used (it's already
close):

1. Find Messages... (there's a menu option for it, but it
doesn't do anything)
2. Insert arbitrary headers
3. Let me add a column to the header-list pane, so I can see
which emails include attachments. And just deal better with
attachments too, like maybe let me save attachments to disk
w/out having to drag/drop to the finder.

C
Post by Malte S. Stretz
As it's free to use the tag as an individual, we (the SA
developers & users)
might support it by adding the tag to our mail.
-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Malte S. Stretz
2002-08-26 21:56:57 UTC
Permalink
Post by Craig R.Hughes
Stupid OSX Mail.app won't let me insert custom headers.
... and oh-so-many other Mailers don't, too (Outlook is just an example).
Post by Craig R.Hughes
There are exactly 3 features missing from this program which would
I disagree; let's start a flame war :-P
Post by Craig R.Hughes
1. Find Messages... (there's a menu option for it, but it
doesn't do anything)
*lol*
Post by Craig R.Hughes
2. Insert arbitrary headers
3. Let me add a column to the header-list pane, so I can see
which emails include attachments. And just deal better with
attachments too, like maybe let me save attachments to disk
w/out having to drag/drop to the finder.
What about saving whole messages eg. to kick them through SA?

Malte
--
-- Coding is art.
--
-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Theo Van Dinter
2002-08-26 22:43:05 UTC
Permalink
Post by Malte S. Stretz
What about saving whole messages eg. to kick them through SA?
He doesn't need SA now that Mail has built-in junk mail capabilities.
Just click the button. :P
--
Randomly Generated Tagline:
"Well Courage, it goes to show ya: if someone takes the time to etch
a poem on a plaque... You'd better take it seriously. Especially if
it rhymes..."
- Muriel on Courage the Cowardly Dog, "The Forbidden Hat of Gold"
Craig R.Hughes
2002-08-26 23:15:06 UTC
Permalink
Post by Theo Van Dinter
Post by Malte S. Stretz
What about saving whole messages eg. to kick them through SA?
He doesn't need SA now that Mail has built-in junk mail capabilities.
Just click the button. :P
Like I'm gonna pay $129 for that when SA works great.

C



-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Tony L. Svanstrom
2002-08-27 00:01:38 UTC
Permalink
Post by Craig R.Hughes
Post by Theo Van Dinter
Post by Malte S. Stretz
What about saving whole messages eg. to kick them through SA?
He doesn't need SA now that Mail has built-in junk mail capabilities.
Just click the button. :P
Like I'm gonna pay $129 for that when SA works great.
129 USD to use Mail.app...?????


/Tony
--
# Per scientiam ad libertatem! // Through knowledge towards freedom! #
# Genom kunskap mot frihet! =*= (c) 1999-2002 ***@svanstrom.com =*= #

perl -e'print$_{$_} for sort%_=`lynx -dump svanstrom.com/t`'



-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Theo Van Dinter
2002-08-27 00:05:10 UTC
Permalink
Post by Tony L. Svanstrom
Post by Craig R.Hughes
Like I'm gonna pay $129 for that when SA works great.
129 USD to use Mail.app...?????
To get the new Mail application you need to upgrade to Jaguar (Mac OS
X 10.2). I guess if that's all you think you're going to get out of it,
that's $129 (actually $119 at most places) for Mail.

I personally like the speed increase and new functionality, but ... ;)
--
Randomly Generated Tagline:
Um, it's like, uh ... did anyone see the movie `Tron'?

-- Homer Simpson
Treehouse of Horror VI
Tony L. Svanstrom
2002-08-27 00:53:24 UTC
Permalink
Post by Theo Van Dinter
Post by Tony L. Svanstrom
Post by Craig R.Hughes
Like I'm gonna pay $129 for that when SA works great.
129 USD to use Mail.app...?????
To get the new Mail application you need to upgrade to Jaguar (Mac OS
X 10.2). I guess if that's all you think you're going to get out of it,
that's $129 (actually $119 at most places) for Mail.
Ah, yes, I was thinking iTo^H^H^H.Mac; my mistake.
Post by Theo Van Dinter
I personally like the speed increase and new functionality, but ... ;)
The speed, yes, but functionality... I don't think the average user will
notice if they're running X.1 or X.2.


/Tony
--
# Per scientiam ad libertatem! // Through knowledge towards freedom! #
# Genom kunskap mot frihet! =*= (c) 1999-2002 ***@svanstrom.com =*= #

perl -e'print$_{$_} for sort%_=`lynx -dump svanstrom.com/t`'



-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Craig R.Hughes
2002-08-27 01:59:29 UTC
Permalink
Post by Tony L. Svanstrom
Post by Theo Van Dinter
I personally like the speed increase and new functionality, but ... ;)
The speed, yes, but functionality... I don't think the average user will
notice if they're running X.1 or X.2.
Am I really going to notice a speed increase over X.1? I've
been severely disappointed in the past with just about every
Apple system upgrade where they promise speed increases (they
always promise speed increases in their system updates), at
least that was my experience back in the ~7.5 days which is when
I last used a mac.

C



-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Tony L. Svanstrom
2002-08-27 02:58:48 UTC
Permalink
Post by Tony L. Svanstrom
Post by Theo Van Dinter
I personally like the speed increase and new functionality, but ... ;)
The speed, yes, but functionality... I don't think the average user will
notice if they're running X.1 or X.2.
Am I really going to notice a speed increase over X.1? I've been severely
disappointed in the past with just about every Apple system upgrade where
they promise speed increases (they always promise speed increases in their
system updates), at least that was my experience back in the ~7.5 days which
is when I last used a mac.
Yes, it pretty much went downhill from 7.5.x, but now we're talking about
what's more or less a new OS, that Apple is finetuning...

I switched almost two weeks ago and I haven't used X.1 since then, so this is
"normal" to me now, but when I updated I felt that the system was a little bit
more responsive (on a non-QE capable PowerBook); esp. the Finder.
Considering that I hardly ever push this computer to its limits I feel that
most people will feel this even more than me.

/Tony
--
# Per scientiam ad libertatem! // Through knowledge towards freedom! #
# Genom kunskap mot frihet! =*= (c) 1999-2002 ***@svanstrom.com =*= #

perl -e'print$_{$_} for sort%_=`lynx -dump svanstrom.com/t`'



-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Theo Van Dinter
2002-08-27 03:08:26 UTC
Permalink
Post by Tony L. Svanstrom
I switched almost two weeks ago and I haven't used X.1 since then, so this is
"normal" to me now, but when I updated I felt that the system was a little bit
more responsive (on a non-QE capable PowerBook); esp. the Finder.
Considering that I hardly ever push this computer to its limits I feel that
most people will feel this even more than me.
Well, I installed this morning and the system does seem much more
responsive (TiBook). I don't necessarily push this guy to the limits
either, but I regularly found the machine sluggish when I was running a
bunch of stuff all together. So far today there haven't been any
problems beyond your standard upgrade issues (some non-standard apps
having library problems, some configuration bits needed tweaking, I
still haven't gotten my dock transparent again...) :)



-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Malte S. Stretz
2002-08-27 09:55:48 UTC
Permalink
Post by Theo Van Dinter
Post by Tony L. Svanstrom
I switched almost two weeks ago and I haven't used X.1 since then, so
this is "normal" to me now, but when I updated I felt that the system
was a little bit more responsive (on a non-QE capable PowerBook); esp.
the Finder. Considering that I hardly ever push this computer to its
limits I feel that most people will feel this even more than me.
Well, I installed this morning and the system does seem much more
responsive (TiBook). I don't necessarily push this guy to the limits
either, but I regularly found the machine sluggish when I was running a
bunch of stuff all together. So far today there haven't been any
problems beyond your standard upgrade issues (some non-standard apps
having library problems, some configuration bits needed tweaking, I
still haven't gotten my dock transparent again...) :)
I read somewhere that they do all the graphics now in the GPU; the whole
desktop is one big 2D-3D scene including bump mapping etc. So the CPU
doesn't have to cope with all that eye-candy...

All AFAIK :)
Malte
--
-- Coding is art.
--
-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Tony L. Svanstrom
2002-08-27 11:47:19 UTC
Permalink
Post by Malte S. Stretz
Post by Theo Van Dinter
Post by Tony L. Svanstrom
I switched almost two weeks ago and I haven't used X.1 since then, so
this is "normal" to me now, but when I updated I felt that the system
was a little bit more responsive (on a non-QE capable PowerBook); esp.
the Finder. Considering that I hardly ever push this computer to its
limits I feel that most people will feel this even more than me.
Well, I installed this morning and the system does seem much more
responsive (TiBook). I don't necessarily push this guy to the limits
either, but I regularly found the machine sluggish when I was running a
bunch of stuff all together. So far today there haven't been any
problems beyond your standard upgrade issues (some non-standard apps
having library problems, some configuration bits needed tweaking, I
still haven't gotten my dock transparent again...) :)
I read somewhere that they do all the graphics now in the GPU; the whole
desktop is one big 2D-3D scene including bump mapping etc. So the CPU
doesn't have to cope with all that eye-candy...
That's Quartz Extreme, what I meant when I wrote that my PowerBook/laptop
wasn't QE-capable; and it's still faster after the update...


/Tony
--
# Per scientiam ad libertatem! // Through knowledge towards freedom! #
# Genom kunskap mot frihet! =*= (c) 1999-2002 ***@svanstrom.com =*= #

perl -e'print$_{$_} for sort%_=`lynx -dump svanstrom.com/t`'



-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Malte S. Stretz
2002-08-27 12:10:04 UTC
Permalink
Post by Tony L. Svanstrom
Post by Malte S. Stretz
Post by Theo Van Dinter
Post by Tony L. Svanstrom
I switched almost two weeks ago and I haven't used X.1 since then,
so this is "normal" to me now, but when I updated I felt that the
system was a little bit more responsive (on a non-QE capable
PowerBook); esp. the Finder. Considering that I hardly ever push
this computer to its limits I feel that most people will feel this
even more than me.
Well, I installed this morning and the system does seem much more
responsive (TiBook). I don't necessarily push this guy to the limits
either, but I regularly found the machine sluggish when I was running
a bunch of stuff all together. So far today there haven't been any
problems beyond your standard upgrade issues (some non-standard apps
having library problems, some configuration bits needed tweaking, I
still haven't gotten my dock transparent again...) :)
I read somewhere that they do all the graphics now in the GPU; the
whole desktop is one big 2D-3D scene including bump mapping etc. So the
CPU doesn't have to cope with all that eye-candy...
That's Quartz Extreme, what I meant when I wrote that my
PowerBook/laptop wasn't QE-capable; and it's still faster after the
update...
Oh. Ok. So that was its name... :o)

Malte
--
-- Coding is art.
--
-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Craig R.Hughes
2002-08-27 01:52:24 UTC
Permalink
$129 to use the Mail.app from Jaguar -- cos you have to buy Jaguar.

C
Post by Tony L. Svanstrom
Post by Craig R.Hughes
Post by Theo Van Dinter
Post by Malte S. Stretz
What about saving whole messages eg. to kick them through SA?
He doesn't need SA now that Mail has built-in junk mail capabilities.
Just click the button. :P
Like I'm gonna pay $129 for that when SA works great.
129 USD to use Mail.app...?????
-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Malte S. Stretz
2002-08-27 09:58:58 UTC
Permalink
Post by Craig R.Hughes
Post by Tony L. Svanstrom
Post by Craig R.Hughes
Post by Theo Van Dinter
Post by Malte S. Stretz
What about saving whole messages eg. to kick them through SA?
He doesn't need SA now that Mail has built-in junk mail capabilities.
Just click the button. :P
Like I'm gonna pay $129 for that when SA works great.
129 USD to use Mail.app...?????
$129 to use the Mail.app from Jaguar -- cos you have to buy Jaguar.
But maybe Jaguar has a working Find Messages... Wouldn't that be an
argument? :o)

Malte
--
-- Coding is art.
--
-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Craig R.Hughes
2002-08-27 15:16:17 UTC
Permalink
Yes, it would be. But I can't find a single thing online which
confirms that you can in fact now Find Messages...

Could one of you 10.2ers fire up Mail.app and let me know?

I'm tempted to sue Apple for false advertising since they have a
"Find messages..." menu option in 10.1 and see if they'll settle
out of court for a copy of 10.2 if that fixes the problem.

C
Post by Malte S. Stretz
Post by Craig R.Hughes
$129 to use the Mail.app from Jaguar -- cos you have to buy Jaguar.
But maybe Jaguar has a working Find Messages... Wouldn't that be an
argument? :o)
-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Theo Van Dinter
2002-08-27 15:27:44 UTC
Permalink
Post by Craig R.Hughes
Could one of you 10.2ers fire up Mail.app and let me know?
Seems to work fine for me. I'm currently searching for the "To" field
in all mailboxes. Takes a little bit, due to amount of mail and the
fact that most things aren't indexed for me (why use Mail.app when I
have mutt?)
--
Randomly Generated Tagline:
On the plus side, it's a lot easier in general to find /usr/include than cpp.
-- Larry Wall in <***@wall.org>
Tony L. Svanstrom
2002-08-27 16:01:52 UTC
Permalink
Post by Theo Van Dinter
Post by Craig R.Hughes
Could one of you 10.2ers fire up Mail.app and let me know?
Seems to work fine for me. I'm currently searching for the "To" field
in all mailboxes. Takes a little bit, due to amount of mail and the
fact that most things aren't indexed for me (why use Mail.app when I
have mutt?)
<AOL>me too</AOL>

Works fine, no problems AFAIK.


/Tony
--
# Per scientiam ad libertatem! // Through knowledge towards freedom! #
# Genom kunskap mot frihet! =*= (c) 1999-2002 ***@svanstrom.com =*= #

perl -e'print$_{$_} for sort%_=`lynx -dump svanstrom.com/t`'



-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Matt Sergeant
2002-08-27 15:34:21 UTC
Permalink
Post by Craig R.Hughes
Yes, it would be. But I can't find a single thing online which
confirms that you can in fact now Find Messages...
Could one of you 10.2ers fire up Mail.app and let me know?
I'm tempted to sue Apple for false advertising since they have a
"Find messages..." menu option in 10.1 and see if they'll settle
out of court for a copy of 10.2 if that fixes the problem.
I've got 6C115 (I think) the current developer release (though not the
same as gold). The "find" thing takes you to the toolbar find box, not
any new find dialog. That find toolbar item allows you to search only on
full text, subject, "to" and "from".

Oh, and I'm not yet 100% impressed with Mail.app's junk filter. It
regularly thinks perl5-porters mail is trash. Oh wait, maybe it's right. ;-)

Matt.




-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Tony L. Svanstrom
2002-08-27 16:31:02 UTC
Permalink
I've got 6C115 (I think) the current developer release (though not the same
as gold).
The only difference AFAIK is the path to a single helpfile, which we don't use
anyways, right? *G*


/Tony
PS I also have a problem with that stupid helpprogram crasching whenever I
search for "windows"; not sure if that's a feature or not. =)
--
# Per scientiam ad libertatem! // Through knowledge towards freedom! #
# Genom kunskap mot frihet! =*= (c) 1999-2002 ***@svanstrom.com =*= #

perl -e'print$_{$_} for sort%_=`lynx -dump svanstrom.com/t`'



-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Craig R.Hughes
2002-08-26 23:09:30 UTC
Permalink
Yeah, unfortunately "Save As..." saves as RTF. Bring me the
product manager's head on a plate.

C
Post by Malte S. Stretz
What about saving whole messages eg. to kick them through SA?
-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Tony L. Svanstrom
2002-08-26 22:28:47 UTC
Permalink
Stupid OSX Mail.app won't let me insert custom headers. There are exactly 3
features missing from this program which would make it the best mail program
2. Insert arbitrary headers
This was a problem with Claris Emailer too, the solution was then to find the
headers in the code and replace it with "your: header\nheader: "; start diging
and with some luck you might find something...


/Tony
--
# Per scientiam ad libertatem! // Through knowledge towards freedom! #
# Genom kunskap mot frihet! =*= (c) 1999-2002 ***@svanstrom.com =*= #

perl -e'print$_{$_} for sort%_=`lynx -dump svanstrom.com/t`'



-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Craig R.Hughes
2002-08-26 21:02:05 UTC
Permalink
Well, there isn't enough data to feed to the GA yet, so we're
having to hand-set scores. If SWE/HIL doesn't pan out, we'll
either remove the rules or change the scores in future
releases. I'm willing to believe that there will be faster
"legitimate" adoption of SWE than illigitimate, and that Habeas'
reponsiveness in at least putting things in the HIL will be good
enough for now. Giving SWE a lower score makes it more or less
useless. It *is* something of an experiment, but I think it's
one that has a decent shot, and so is worth trying.

C
Post by Daniel Quinlan
I don't know why so many people are buying into the Habeas marketing
material and assigning huge scores to the mark before we see whether
it will actually work. I think the HIL is a better bet, but only
because we've seen that DNSBLs work well enough to use.
I think it's an interesting idea and I hope it works to reduce spam,
but there is no reason to project hope into large SA scores. We might
as well dump the GA if that's how we're going to operate.
Dan
-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Justin Mason
2002-08-26 15:46:21 UTC
Permalink
Post by Craig R.Hughes
The GA run on the current corpus is yielding an average (mean)
score for false-positives of about 10 points, which is higher
than before. There's some tweaking to be done yet (rule
elimination and such), so that will likely fall, but -15 might
not be enough of a correction in many situations. I will set
the scores for the Habeas stuff and the Ironport stuff to be
a) estimated likeliness of compliance with Habeas/Bonded sender
rules, and
b) score correction level needed to achieve purposes of those projects.
Using the score-ranges stuff and my aggressive anti-FP evaluation code,
BTW, yielded an average FP of 5.7 points, with 0.67% FPs and 96.57%
overall accuracy -- but I notice you went back to the other method.
Just pointing this out ;)

--j.


-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Craig R.Hughes
2002-08-26 16:02:07 UTC
Permalink
Here's where I ended up overnight:

# SUMMARY for threshold 5.0:
# Correctly non-spam: 153975 68.08% (99.64% of non-spam corpus)
# Correctly spam: 64893 28.69% (90.57% of spam corpus)
# False positives: 550 0.24% (0.36% of nonspam, 25614 weighted)
# False negatives: 6756 2.99% (9.43% of spam, 4903 weighted)
# Average score for spam: 19.4 nonspam: -12.0
# Average for false-pos: 10.0 false-neg: 0.7
# TOTAL: 226174 100.00%

So I guess it turns out my code is more anti-FP aggressive after all :)

Justin, I'm in meetings for a while this morning, so I'll send
you over my scores file. If you could check the problem in
logs-to-c which seems to have overlooked a couple of rules, then
we can start carving rules out based on the results of
post-ga-analysis (which is a great way of seeing which rules are
helping/hurting in terms of creating/revoking false postives and
false negatives. In my opinion, the post-ga-analysis gives much
better insight for rule revocation than does the
hit-frequencies. Will send newscores and analysis under
separate email.

C
Post by Justin Mason
Post by Craig R.Hughes
The GA run on the current corpus is yielding an average (mean)
score for false-positives of about 10 points, which is higher
than before. There's some tweaking to be done yet (rule
elimination and such), so that will likely fall, but -15 might
not be enough of a correction in many situations. I will set
the scores for the Habeas stuff and the Ironport stuff to be
a) estimated likeliness of compliance with Habeas/Bonded sender
rules, and
b) score correction level needed to achieve purposes of those
projects.
Using the score-ranges stuff and my aggressive anti-FP evaluation code,
BTW, yielded an average FP of 5.7 points, with 0.67% FPs and 96.57%
overall accuracy -- but I notice you went back to the other method.
Just pointing this out ;)
--j.
-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Justin Mason
2002-08-26 17:01:52 UTC
Permalink
Post by Craig R.Hughes
you over my scores file. If you could check the problem in
logs-to-c which seems to have overlooked a couple of rules,
fixed -- it wasn't counting the new "meta" rules.
Post by Craig R.Hughes
then we can start carving rules out based on the results of
post-ga-analysis (which is a great way of seeing which rules are
helping/hurting in terms of creating/revoking false postives and false
negatives. In my opinion, the post-ga-analysis gives much better
insight for rule revocation than does the hit-frequencies.
OK, I haven't seen it yet, but it looks good. (I reimplemented that wheel
slightly btw in hit-frequencies, but let's just use post-ga-analysis
instead ;)

--j.


-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Craig R.Hughes
2002-08-26 18:09:41 UTC
Permalink
Awesome, I didn't realize meta rules were in for 2.40! I was
planning on tweaking the GA for meta rules to have it create its
own meta rules... I guess that'll have to wait for 2.41/2.50

We should probably at least stick a meta rule in for
HABEAS_SWE/HABEAS_HIL

So the GA should probably get re-run then to include handling
those meta rules?

C
Post by Justin Mason
Post by Craig R.Hughes
you over my scores file. If you could check the problem in
logs-to-c which seems to have overlooked a couple of rules,
fixed -- it wasn't counting the new "meta" rules.
-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Justin Mason
2002-08-27 10:17:24 UTC
Permalink
Post by Craig R.Hughes
Awesome, I didn't realize meta rules were in for 2.40! I was
planning on tweaking the GA for meta rules to have it create its
own meta rules... I guess that'll have to wait for 2.41/2.50
Well, the decision trees stuff is apparently excellent in this way, and I
think that should be well underway at the moment; I'd see how that's
going, rather than trying to work out how to cajole a perceptron into
doing this.

I think, if we do wind up doing another release of 2.xx, we could probably
use data from the decision-tree output to come up with a few meta tests,
to cover the most common cases.
Post by Craig R.Hughes
We should probably at least stick a meta rule in for
HABEAS_SWE/HABEAS_HIL
Good idea -- but not in 2.40 release branch, in HEAD!
Post by Craig R.Hughes
So the GA should probably get re-run then to include handling
those meta rules?
yes. :( I presume you already restarted it last night ;)

--j.


-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Craig R.Hughes
2002-08-27 15:22:19 UTC
Permalink
Post by Justin Mason
Post by Craig R.Hughes
Awesome, I didn't realize meta rules were in for 2.40! I was
planning on tweaking the GA for meta rules to have it create its
own meta rules... I guess that'll have to wait for 2.41/2.50
Well, the decision trees stuff is apparently excellent in this
way, and I
think that should be well underway at the moment; I'd see how that's
going, rather than trying to work out how to cajole a perceptron into
doing this.
It wouldn't really be cajoling a perceptron; I'd extend the GA
to allow the genome length to grow and shrink. It could shrink
by just discarding any rule, grow by creating new meta rules.
Since it could build meta-meta rules if it were so inclined, and
the meta-rule-building/score-allocation would be part of the
same evolutionary process, it might actually yield slightly
better results than a decision tree which separated those 2
parts. I haven't though about it in detail though; it's
possible that it would essentially be building a tree by
stumbling around in the dark, which would be a pretty dumb way
of doing it :)
Post by Justin Mason
I think, if we do wind up doing another release of 2.xx, we
could probably
use data from the decision-tree output to come up with a few
meta tests,
to cover the most common cases.
Another benefit of allowing the genome length to wander would be
that the GA would delete obsolete rules all by itself :)
Post by Justin Mason
Post by Craig R.Hughes
We should probably at least stick a meta rule in for
HABEAS_SWE/HABEAS_HIL
Good idea -- but not in 2.40 release branch, in HEAD!
Yes, yes. Although actually since it's hand scored, and not
evolved, and not subject to mass-check (since the meta rule
depends on a net check), we *could* add it on b2_4_0 :)
Post by Justin Mason
Post by Craig R.Hughes
So the GA should probably get re-run then to include handling
those meta rules?
yes. :( I presume you already restarted it last night ;)
Yup. Let me check on it now...

Ooh, it's done.

# SUMMARY for threshold 5.0:
# Correctly non-spam: 146527 67.40% (99.69% of non-spam corpus)
# Correctly spam: 64524 29.68% (91.65% of spam corpus)
# False positives: 456 0.21% (0.31% of nonspam, 19367 weighted)
# False negatives: 5881 2.71% (8.35% of spam, 5366 weighted)
# Average score for spam: 18.3 nonspam: -10.7
# Average for false-pos: 8.9 false-neg: 0.9
# TOTAL: 217388 100.00%


Not too bad. I'll forward the newscores and analysis to you now.

C



-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Justin Mason
2002-09-03 16:30:11 UTC
Permalink
Doh! Why did someone do that?
I cannot tell a lie, 'twas me. The idea was to use the most efficient
databases if they were installed. So what to do?

- document that any existing AWL will become obsolete and be
regenerated from scratch, or

- add automatic-migration code (a total PITA and portability nightmare)
or

- revert back to old code (easy ;). I don't think we really *need* a
massively efficient DB for the AWL.

Votes?

--j.


-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
Matt Sergeant
2002-09-03 16:33:13 UTC
Permalink
Post by Justin Mason
Doh! Why did someone do that?
I cannot tell a lie, 'twas me. The idea was to use the most efficient
databases if they were installed. So what to do?
- document that any existing AWL will become obsolete and be
regenerated from scratch, or
- add automatic-migration code (a total PITA and portability nightmare)
or
- revert back to old code (easy ;). I don't think we really *need* a
massively efficient DB for the AWL.
Votes?
I'd migrate back to the old code. Major stuff like changing the backend
DB should be saved for 3.0.

Sorry I didn't pipe up earlier, but I don't use the AWL at all.

Matt.



-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390

Loading...