| LED Digest 1744: Joe Jobs, Open Source Bayesian, and more! |
|
|
|
================================================== The LED Digest Moderated Discussion List "Effective Online Advertising, Since 1997" pair Networks: The LED's Web Host Hosting and Domain Reg. from a Trusted Leader pair.com for Hosting | pairNIC.com for Domains ================================================== List Moderator: Published by: Adam Audette LED Digest adam,led-digest.com http://www.led-digest.com ................................................ February 9, 2004 Issue #1744 ................................................ .....IN THIS DIGEST..... ==== CONTINUING ================= --== Spammer Using our Domain, Help! ==-- ~ Dave Starr "Site Point just published a comprehensive and useful article on this very problem." ~ Daniel Lance Herrick "That is called a "Joe Job" after the first victim, Joe Doll." --== Losing Filtered Email ==-- ~ Adam Boettiger "The key has always been 'How do I reduce spam without a high number of false positives?'..." ~ Brett Dorron "Rod, your solution is akin to stalking, and is no different from attaching secret tracking devices to your products..." --== Email Address Verification ==-- ~ Tamra Heathershaw-Hart "I'm one of the folks who lie on every single one of those forms." ~ Wanda Husick "This is not a technical answer to your question, but it may work anyway..." ~ Malcolm Bailey "I discovered a way to verify email addresses on the fly with a high level of accuracy." ==== BILLBOARD =================== --== Bayesian Filtering for Outlook ==-- ~ Russ Swall <Moderator Comment> ===== CONTINUING ================================= From: Dave Starr Subject: Joe Job [was Spam-spoofed] > Some unscrupulous person overseas has been using > our email addresses to send tons and tons of spam all > over the internet. I am getting the bounces by the dozens > (sometimes hundreds). - Brigitte Burchett, LED 1743 Brigitte, Sorry to say it, but you seem to have fallen victim to a "Joe Job", so-called because it first came to internet attention in a noticeable way via a massive attack on www.joes.com back in 1997. Fortunately, Site Point just published a comprehensive and useful article on this very problem. It won't be an easy battle, but at least this article gives clear-cut, actionable steps to take. http://www.sitepoint.com/article/1281 Best regards Dave Starr www.retiredpay.com ------- new post - same topic ------- From: Daniel Lance Herrick Subject: Spam spoofed (Joe Job) The first thing we have to say is that essentially all the recent Microsoft infecting viruses that send e-mail to everyone in the addressbook on the infected computer select From addresses from the same place. It could just be that a bunch of people who have you in their addressbooks are infected by recent viruses. However, you may well be right that someone has maliciously chosen you as a target. That is called a "Joe Job" after the first victim, Joe Doll. You can read an account of his story at http://www.everything2.com/index.pl?node=Joe%20Job , his own brief description is at http://www.joes.com/spammed.html Another victim offers a longer story at http://members.cox.net/joejob/ There are a couple who did something about it at http://www.techdirt.com/commentShow.pl?sid=20030211/0230225&pid=22 And, of course, a Google search for "joe job" (use the quotation marks to filter out joes and jobs) turns up a lot more. Didn't lead me to a rather detailed description of how to follow the things being advertised and shut them down that I read the last time I was chasing the term "joe job". Adding track and down to the search turned up some useful and helpful pages. Here is one story of a successful search for such a perpetrator: http://www.wired.com/news/print/0,1294,60635,00.html There may be some help in those pages, there should be some comfort that you are not alone. Finding them and sueing for defamation of character and related offenses would be a useful contribution. daniel lance herrick ------- new post - new topic -------- From: Adam Boettiger Subject: Losing email > Instead of looking at all the "noise" in a mailbox... why not > look for only the "signal among the noise" - the truly wanted > email? You can easily do this if you use a whitelist-based > solution... - Joe Halbrook, LED 1743 Joe is, of course, correct that whitelisting is one of the most effective methods of ensuring email does get through. And the site form tactic is a good one. But don't give up on Bayesian filters just yet. I used to think that spammers putting all the garbage and odd words before or hidden in their ad worked to trick the filters too. And for some products it does that. But for others it does not. The key has always been "How do I reduce spam without a high number of false positives?" (real email classified as spam and filtered out). There are many ways to do it but one worth exploring is an open source project called SpamBayes. For the novice user it probably is not the best solution, but if you know your way around Outlook fairly well and can install programs, this is one filter that teaches itself and does not get tricked by the extraneous crap that it in spam. http://spambayes.sourceforge.net/ It's a python-based program that has a plug-in for Outlook. It takes some time to install and you have to print the directions and follow them to a T, but once set up it is very cool. You start out by creating two folders in Outlook: "Ham" and "Spam" - funny. Then you go through your Inbox or Deleted items folder and drag or move about 40-50 messages into one folder or the other to classify them. The more you start with, the faster it "learns". Then you click the Teach button and it analyzes them and from that point on will filter about 95% of spam effectively to your Junk Mail folder with maybe 1 or 2 false positives. You have the option of telling it that anyone in your address book already is not spam. But as you go, if you get one in your Inbox that is spam, there's a plugin button on Outlook that allows you to classify it as spam and the program will learn from the message - not the crap words at the begining but actual data points - and I've gotten mine down to about 99% effectiveness. If it is not sure if something is spam or not it will put it in a folder that is called Junk Suspects and I go through that once a week and reclassify what's in there. Thus far beyond whitelisting and spamarrest.com, this has been the best filter I've seen, but again, since it is open source it does require a learning curve to install. But the nice thing is, it's also free and continually being worked on by many developers because of the open source status. HTH Adam Boettiger, President I-Advertising - Interactive Advertising & Marketing Since 1996 http://www.i-advertising.com ------- new post - same topic ------- From: Brett Dorron Subject: Losing email > We use the program to embed the hidden mode > code (it can be visible also, but I seem to enjoy the > stealth mode where I can voyeur their interest in my email). - Rod Aries, LED 1742 I've been a proud recipient of LED for some years now, but today, when I read through Rod's comment, I felt ashamed to be associated. Rod, your solution is akin to stalking, and is no different from attaching secret tracking devices to your products, or having someone secretly follow your customers home from your store. I'd guess the greatest proponents of this type of tracking would be the spammers, and I would hope there are, or will be ways to counteract this practice. I am too shocked to type... I'll keep an eye out for your sites. I also believe that the greatest reason for fake email addresses in web forms, is the fear of spammers, and trackers. I fake it everytime I can, and often go elsewhere when I can't. I get about 1 spam mail per month. Brett Dorron ------- new post - new topic ------- From: Tamra Heathershaw-Hart Subject: Email verification > Before you are allowed to download any of the information > you have to fill out a short form consisting of first name, last > name, telephone number, email address and country. We > get a lot of people typing... bogus email addresses. > Is there a method of verifying the email address on the fly...? - Andy Johnson, LED 1740 I'll leave it to some other member of the list to rant about how no brick-and-mortar establishments take names and require telephone numbers before they'll talk about their products so wanting a programmatical way to guarantee that the email is legit just isn't good business. (Shoot, you can even test drive a new $50,000 automobile without first giving the sales folks as much detail as some of those forms ask for just to see a crummy PDF.) Instead, I'll tell you about life on the other side. You see, I'm one of the folks who lie on every single one of those forms. I go through a lot of product demos, look at a lot of downloadable brochures, and I NEVER give my real name and only use a throw-away email address (or fake email addresses if an email reply isn't needed to access the form). Why? Because I most adamantly DO NOT WANT the sales folks calling me (or emailing me) and that's what always happens if you don't lie on the forms. Am I a hater of telemarketing or rabid anti-spammer? Uh, no, actually. I have no problem talking to or reading email from folks who are selling things I want or need. I'm a marketing designer and the only reason I'm looking at those demos and etc. is to check out my clients' competition. I have NO desire to buy the stuff that's being sold... but I do have an interest in seeing the brochure, and in trying out the demo product interface, and definitely in looking at any product demos. It's called competitive research, and it's one of the things that makes for really great graphic design, because knowing what the competition is doing helps me design things that are different yet directly compete in terms of quality. (And I hate crowds so I'm much happier not doing this type of market research at trade shows.) Unfortunately for you form-creating folks, I'm also VERY tired of explaining to quota-driven sales folks that I'm not a potential customer and have no interest in their products. So, to avoid explaining to yet another poor schmuck selling JAVA software, SONET equipment, or water treatment systems for chip manufacturing that I'm just a partner in a small design firm and would have no interest in their wares even if I had a zillion dollars in the bank... I lie on the forms. Maybe if those forms had a nice little checkbox for "Are you a graphic designer or marketing person doing competitive research?" I'd use my real name. But until some form field that basically says "don't call / email me I'm just looking" is standard (and taken seriously) on all sites wanting detailed info just to look at a PDF brochure expect to see plenty more forms filled out by Ted E. Bear, Gold E. Locks, and their many friends. Tamra Heathershaw-Hart Crendo = Create + Enhance + Deliver www.crendo.com ------- new post - same topic ------- From: Wanda Husick Subject: Email verification This is not a technical answer to your question, but it may work anyway. Perhaps if you guaranteed people you were not going to sell their email address or use it for sending spam, then they will use their real address. Tell them why you want their address (such as to follow through with selling your own product). If you cannot guarantee people that putting their email address on a form is not going to result in an inundation of spam email - then I do not blame them for putting false email addresses down. Wanda Husick ------- new post - same topic ------- From: Malcolm Bailey Subject: Email Address Verification > Is there a method of verifying the email address on the fly...? - Andy Johnson, LED 1740 > We think that the answer is no, you cannot. > If anyone knows different, we'd love to know:-) - Ian Dickson, LED 1742 Hi All and Thanks Adam for keeping the list great! A while ago I was doing some research into this and discovered a way to verify email addresses on the fly with a high level of accuracy. The method will never reject a valid email address, but will reject invalid addresses 90% of the time. I've tested it and built it into (well OK hacked it into) a small scale system for my own use and have always been tempted to develop it further, into a product for others to use. Before I begin development I need a bit of market research: 1) What exactly would people want? (An ASP model Vs License and install on their own server) 2) Would a confidence result (0,20,40,60,80,100%) be beneficial? 3) What would people be prepared to pay? e.g. ?10 Per 1000 emails validated or ?50 one off fee? (these are figures just plucked from the air - feel free to suggest what value it would add to you!) 4) Are there any periphery services that you'd like to see with it? (bulk validation of an uploaded list etc) 5) What security concerns do people have? I'd love to get peoples thoughts and feedback - please send to: This email address is being protected from spam bots, you need Javascript enabled to view it If there appears to be some interest in this I'll hopefully get around to it in the near future! Cheers, Malcolm Bailey ==== BILLBOARD ==================================== From: Russ Swall Subject: Bayesian filtering > ... I suggest you look into any of the many products now > available that use Bayesian filtering to identify spam. I use > Eudora, which has it built in. - Veronica Yuill, LED 1742 Does anyone have recent experience with Bayesian filtering in conjunction with Outlook 2003. I have attempted (a couple of times) to find a good Bayesian filter, but have so far been unsuccessful. BTW, OL 2003 spam filtering is pretty good, but the main problem remains: you still have to sift through the junk in case a good one got "junked." Russ Swall Vision Data Solutions http://visionds.com <Moderator Comment> Russ, see Adam Boettiger's post above (under the thread "Losing Email" - 3rd post from the top). He's got a great solution with an open source bayesian plug-in for Outlook. Check it out. -adam ------------------------------------------------------- The LED Digest is sponsored by pair Networks: pair.com for Hosting | pairNIC.com for Domains Copyright 1995-2004 Adam Audette. All Rights Reserved. ----------------------------------------------------------------- "The only true wisdom is in knowing you know nothing." - Socrates |
