Home

Bio & Disclosures

Discussions


xFruits

2007 Events

Wednesday, April 28, 2004

Previous topic
Next topic
inactiveTopic Wednesday, April 28, 2004
started 4/28/2004; 4:40:56 AM - last post 4/30/2004; 7:20:24 AM
Doc Searls - Wednesday, April 28, 2004  blueArrow
4/28/2004; 8:40:56 AM (reads: 5266, responses: 4)
Are (or Eire) you sure he's not here? 
 The first two Irish blogs I read both mention Dave.
 Speaking of which, how about a get-together here? Anybody up for that? (I am, like, not organized, as you can tell.)
 
KOAF! 
 Say happy birthday to Bilal Hasan Marti. And congrats to his Dad and mom.
 Bonus link: Don's combination case-for-FOAF and all-purpose reply to social network site invitations.
 
We can ass-check your facts, too 
 Mike Rodriguez takes issue with the choice of photos on the cover of the new U.S. News & World Report. In it a left-facing red-hued young John Kerry looks like Che Guevara in a tie (his hair is its own beret), while true-blue uniformed George W. Bush looks about as pre-presidential as can be. Sez Mike:
 Compare that to the photo of John Kerry, clad in a suit, on the left (oh! I just noticed that one!). I told myself that, yes, it's biased, but, and this is where I give them credit, it's true. In 1971, Kerry was already out of the service, and Bush was nominally serving his country.
 Here's the cover story that goes with the picture.
 
Black chopper notification 
 Sean Bonner: Feds reading blogs.
 
Worker be 
 Frank has an important question for all contributors to journals like IT Garage:
 Here's the challenge: what can I write about without giving away the store from my customers' points of view? I have two ideas..
 By the way, Frank is the first IT mechanic (trust me, I don't qualify... I just run the place), to start plying his craft at IT Garage.
 
Assisting an assassin 
 I have a friend whose mail always gets flagged as spam because... well, I'll let SpamAssassin tell you:
 This mail is probably spam....Content preview:

Content analysis details: (7.70 points, 5 required)
HTML_FONT_FACE_ODD (0.3 points) BODY: HTML font face is not a commonly used face
HTML_FONT_BIG_B (0.5 points) BODY: HTML has a big "font" and "B" tag combo
HTML_50_60 (0.5 points) BODY: Message is 50% to 60% HTML
HTML_MESSAGE (0.1 points) BODY: HTML included in message
HTML_FONT_BIG (0.3 points) BODY: FONT Size +2 and up or 3 and up
HTML_FONT_COLOR_BLUE (0.1 points) BODY: HTML font color is blue
HTML_IMAGE_ONLY_08 (0.7 points) BODY: HTML has images with 600-800 bytes of words
SUBJ_ALL_CAPS (1.1 points) Subject is all capitals
FORGED_MUA_OUTLOOK (3.5 points) Forged mail pretending to be from MS Outlook
MISSING_OUTLOOK_NAME (0.6 points) Message looks like Outlook, but isn't

The original message did not contain plain text, and may be unsafe to open with some email clients; in particular, it may contain a virus, or confirm that your address can receive spam. If you wish to view it, it may be safer to save it to a file and open it with an editor.

 He's using generic Outlook on XP, he believes. What should I (or we) tell him to do? His business requires sticking with Windows and Outlook, so snarky entreaties to ditch either won't be useful.
 
NEAT stuff 
 We got up early for two reasons Monday morning: 1) so I could catch a bus to LAX to catch a plane I thought was at 10:55am but turned out to be at 5:50pm; and 2) to see Comet Bradfield. Conditions in the sky turned out to be just as hazy as those in my mind, so we were 0 for 2.
 Still, whether or not many of us catch Bradfield, there's still the prospect of NEAT C/2001 Q4, which is just one of several cometary discoveries produced by the Near Earth Asteroid Tracking Program.
 Back when I was in Top 40 radio, there was still an art to "smacking a lyric" by talking over the instrumental intro to a song. I've just done that for the latest post by Michael Taht, who has much more to say about most of the above. Hit it, Mike.

discuss

Grant - Re: Assisting an assassin  blueArrow
4/28/2004; 9:50:03 AM (reads: 574, responses: 0)
The easiest thing to do would be to encode the email in plain text. Sure, he wouldn't be able to do all the bold and colors and big fonts, but also his emails won't be flagged as spam. It's a tradeoff I suppose.

To change this you must go to Tools -> Options -> Mail Format -> Compose in this message format: Plain Text.

Of course, that only gets us part way there. Taking out all the points for HTML stuff, you are still left with the 1.1 for all caps in the subject, 3.5 for "Forged mail pretending to be from MS Outlook" and .6 for "Message looks like Outlook, but isn't" which all still adds up to more than 5.

To get the score below five all your friend must do is not put his subjects in caps. But that still leaves that 4.1 points that I don't understand at all. Hopefully somebody else will be able to take it from there, but at least this way SpamAssassin won't flag the emails as spam.

discuss

Lon Baker - Re: Assisting an assassin  blueArrow
4/28/2004; 11:32:40 AM (reads: 567, responses: 0)
In addition to the previous suggestion of changing from html to plain text e-mail format.

You can train SA that your friends e-mail is not spam but ham instead, as well as enable auto-whitelisting to help with this problem.

To teach SA an email is no spam you can:

sa-learn --showdots --ham --dir directory

The directory can contain messages you know are not spam. This will teach the SA bayes filter. You can also reverse this process to teach it spam that gets through SA. SA is smart enough to know about its own tags so you do not have to strip them before training SA in this manner.

The auto-whitelisting is a feature which adds counter weight to known good sources as they successfully send message to you. If an e-mail from a good source happens to breach the SA score it will be adjusted based on the built up whitelist score.

discuss

Rob Schneider - Re: Wednesday, April 28, 2004  blueArrow
4/28/2004; 1:18:04 PM (reads: 589, responses: 1)
Your friend is sending mail in HTML format and is doing way too much formatting. This formatting is either deliberately applied or controlled by the template (see below--my hunch is Word is the editor).

To completely avoid, use "plain text". There are no special settings in Outlook which create complicated HTML. Unless absolutely required (to communicate as opposed to impress), your friend should avoid using HTML formatted mail.

I suspect your friend is using Word as the Editor for making this mail. Just a hunch. Word and Outlook are pretty good together as a tool for writing. I use this combination all the time. My hunch is that your friend's normal.dot template (on which the email document is based, has got complex formatting, and you friend is a lot of different styles, or hand-coding these complex formats ... and SpamAssassin is detecting it.

Unless your friend is required to keep normal.dot, just delete it and start afresh. (Word re-creates it on next startup.) Or have your friend find out what template is being used if not normal.dot.

If your friend wishes to pursue this, your friend can get more advice from Microsoft's newsgroup on Outlook by asking questions about how HTML is format is controlled when using Word as the Editor and composing mail in Outlook.

Finally I was curious about the last two SpamAssassin tests. I looked at http://www.spamassassin.org/tests.html which list all the standard tests. I can't find the tests entitled: FORGED_MUA_OUTLOOK and MISSING_OUTLOOK_NAME. Where did they come from? I used Google to search for these two rules and got lots of hits. I don't know enough about SpamAssassin to really understand the results. Since these seem to be non-standard rules, I'm wondering how they could be so widely distributed so that your friend's mail is "always" detected as spam by SpamAssassin. Something doesn't seem right here--just a hunch.

discuss

Rob Schneider - Re: Wednesday, April 28, 2004  blueArrow
4/30/2004; 11:20:24 AM (reads: 622, responses: 0)
To complete the thread, someone emailed with info on the Spamassassin rules:

"Old(er) versions of Spamassassin used to have lots of rules regarding things like email user agent strings. The spammers realised that these could be used to make a mail look *more* legitimate and thus those rules have been abandoned in spamassassin's greater than 2.55 (thus the reason they arren't on the web)..."

discuss




Copyright 2009 The Doc Searls Weblog

Membership : Join Now : Login

Create your own Manila site in minutes. Everyone's doing it!

Click to see the XML version of this web page.

Blogroll

 
Search archives

Santa Barbarians
Edhat
SB Independent
SB Newsroom
Kevin Barron
Blogabarbara
Craig Smith
SB*Free Press
Joe Andieu
Patrick Gregston
John Quiimby
Das Williams' dad
Katy Pearce
Taymar Pixley
Lisa Gates
Cookie Jill

Everybody else
Spot-on
RageBoy
MysticBourgeoisie
David Weinberger
Miscellaneous
Dave
Berkman
John Palfrey
IT Garage
Bret Fausett
Susan Crawford
Bruce Sterling
Steve Lewis/Bubkes
Hak Pak Sak
Brad Kava
Brad Templeton
Sheila Lennon
Don Marti
Steve Urquhart
Wes Felter
Brad DeLong
Tom Evslin
Brian Oberkirch
Dean Landsman
Hugh MacLeod
LAist
Jeremy Ruston
Geoff Jones
Vaspers the Grate
Sig Rinde
Chris Albritton
Ronni Bennett
Thomas Hawk
Kevin Bedell
Howard
Bryan
Deep Fun
BoingBoing
edhat
Terry Heaton
Jay Rosen
Kim Cameron
George Lakoff
Scott Rosenberg
Larry Lessig
Jim Thompson
Jeff Jarvis
David Isenberg
Stephen Johnson
Tim Oren
Geoff Moore
Rex Hammock
This is Broken
Max Sawicky
Stuart Hughes
Dave Pentecost
John Perry Barlow
Mary Hodder
Dan Gillmor
Steve Gillmor
Dean Landsman
John Stodder
Seth Finkelstein
Renee Blodgett
misbehaving.net
Ruby Sinreich
Ed Cone
Julie Leung
Ted Leung
Ken Coar
Flemming Funch
Mike Sanders
Marc Canter
Joi Ito
Ethan Zuckerman
Doug Kaye
Jon Lebkowski
Judith Meskill
Allen Searls
Esther Dyson
Christopher Lydon
Russell Beattie
Tim Bray
Brian Millar
Mark Pilgrim
Michael Hall
Backup Brain
Frankston, Reed
Britt Blaser
Brent Simmons
Loic Le Meur
Leslie Winer
Mike Taht
Eric Raymond
Volokh Conspiracy
Steven Levy
Lisa Rein
Skywave
Epeus' epigone
Glenn Reynolds
James Taranto
Frank Paynter
Ross Mayfield
Dana Blankenhorn
Ken Bereskin/Panther
Daily Wireless
Filchyboy
OxBlog
Bryan Field-Elliot
Rajesh Jain
Oliver Willis
Gary Turner
Michael O'Connor Clarke
Jennifer Balderama
Kevin Werbach
Amy Wohl
Phil Windley
Fulcrum
Real Joe
Greater Democracy
Mitch Ratcliffe /biz
Mitch Ratcliffe/soc
Wayne Robins
VivaCapitalism
Cut on the bias
Howard Greenstein
The Poor Man
Mickey Kaus
Dave Sifry
Buzz Bruggeman
Ben Hammersley
Matt Jones
Paul Andrews
John Robb
Schoolblog
Tom Shugart
Matt Welch
Blur Circle
Denise Howell
JY
BlackHoleBrain
Chris Pirillo
Marek
Tony Pierce
Chris Nolan's
Spot On

Wil Wheaton
Meg
Brian Linse
Dan Pink
Dawn Olsen
Craig
Yoz
The Head Lemur
Ev
Jeremy Zawodny
Susan Kitchens
K5
Anu Gupta
Jonathon
Fishrush
Dave Ely
Euan Semple
Eric Norlin
Paul Boutin
James Lileks
David Williams
Mary Wehmeier
Bruner Blog
Halley Suitt
Webword
Ann Salisbury
Om Malik
Moxie
J's Notes
Meesh
NUblog
TBTF
Cam
Seth Finkelstein
Tom Matrullo
Chip Hoagland
Deborah
Fortboise
J.D. Lasica
Photodude
Phil Wolff
Andre Durand
Eric Hansen
Mike McBride
Jeneane Sessum
Chris Nolan
Gonzo Engaged
Michael Mussington
UseTheSource
Wes
Adam
Sam Ruby
Miguel
Frank Field
Rebecca Blood
Joshua Allen
Cluetrain
JOHO
EGR
Searls site
Scoble
AKMA
Kottke
Tomalak's Realm
Tim O'Reilly
Mitch Kapor
Bill Quick
Dan Bricklin
Lou Josephs
Alan Reiter
N.Z. Bear
Todd Morman
Zeldman
Glenn
Joshua
Rex Hammock
Matthew Thomas
Brian Dear
Baylink
Burningbird