Thursday, 15 June 2006, 18:52:02 EDT

Public service announcement:

If you get the following email, delete it:

IRS Scam

The "click here" link goes to "http://mail.halas.com.cn/.../IRS/refund/caseid886432/index.html". Clearly not irs.gov. In fact, from irs.gov/privacy/: "Do not respond to unsolicited e-mails that claim to come from the IRS, or any e-mail from an unknown party asking you to submit personal, tax, or financial data. The Internal Revenue Service does not use e-mail to request this type of information."

Categories:

Or use Mozilla Thunderbird. It will detect that the email contains fraudlent addresses and mark it as junk :)

Posted by Mr Frosti on Sunday, 25 June 2006, 22:27:09 EDT.

A couple of things.

1) This post was not meant for the technical crowd that read this site. It was intended for those who don't know a lot about this type of stuff and may not be able to spot the scam. There are a few people like that who read this site.

2) I don't particularly like Thunderbird. When I was managing my email with my home machine, that is, all my mail was delivered directly to my computer with no intermediary POP server, I preferred to use Sylpheed-Claws. It is a much better email client than Thunderbird ever will be. Actually, by the time I had moved to gmail I was using mutt as my client because I had broken something with Sylpheed and didn't feel like trying to fix it. I still like Sylpheed better, though.

3) Thunderbird's junk filter still need to be trained. Just like every other email client with adaptive filtering. True, it probably would have picked this one up. But I would have expected gmail to pick it up, too. However, I am not surprised that it didn't. I currently have 1,363 spam that have (mostly) been automatically picked up by the gmail filter; still, there are a few messages, out of the hundred or so, that I receive every day that I need to manually mark as spam. A Bayesian based filter can do a pretty good job, but it is still just an algorithm and won't catch everything.

Posted by James Sumners on Monday, 26 June 2006, 7:58:56 EDT.

Sylpheed is pretty good from what I remember. It is just ugly. I know it seems like a trivial point, but if I have to stare at it all day, it needs to feel a little less like I am hurting myself.

And as far as Bayesian filters go, mine is nearly human. It has changed my ideas on what I thought was possible for a trained junk filter to catch. I can't even understand how people can use Outlook and its "High" "Medium" or "Low" junk mail settings and be content. We have to supplement the demand for filtering on campus with SpamAssassin. Maybe Outlook 2007 will get things right - but I doubt it.

Posted by Mr Frosti on Friday, 21 July 2006, 22:53:45 EDT.