Wednesday, February 28, 2007

English Spams

I receive a lot of spams in both English and Japanese every day.

There is an interesting difference between English spams and Japanese ones
recently.

In case of Japanese spams, usually they are written in conventional text or html body parts and most of them are trying to lead me to meet-a-mate kind of sites.

I'm feeling that English spams are more advanced in terms of breaking spam filtering.

Here is an excerption from a spam that I received today.

|Subject: linked or
|
| Title was invalid empty an linked! Was invalid empty an linked?
| Main from special toolssign.
| An linked or interwiki contained, main. Invalid empty an linked or?
| Title was invalid empty an linked, or interwiki.
| Error bad titlefrom wikipedia, the free to navigation searchthe. Bad
[snip]
| Titlefrom wikipedia the, free.
| Special, toolssign in create.
| An linked or interwiki contained main from special.

In short, the above lines are nonsense.

At first, looking at the first several words, I thought this might be a linkedIn releated notification or something.

But, the below image was the true body part of the spam. :o
What I thought cool is that it's not only an Ad. just simply written
in a graphic image format but the lines are on the angle.
I'm wondering if there are effective text mining(?) tools which can
retrieve text information with such kind of (in the very near future, non-linear) transformation.

Image:

No comments: