SPAM MAILS: Someone helpd me out here !!
 

... I've been receiving lots of these spam mail messages lately, it just looks like random text, what's the story behind that? are they just testing deliverables ??

SAMPLE:

----ALT--LMZE80393946701088
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 8bit

auditor phytoplankton tenon sticky dixie precede precursor crankshaft crosby swimsuit
auction periphery accelerate redundant videotape architectural footfall periphery trisyllable arum
equidistant theist spontaneous canvasback clean convulsive whereon ferrer espouse guesswork

----ALT--LMZE80393946701088
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 8bit

It's subliminal. If you read it, you will kill your neighbor tonight.

Must KILL :ak47:

They're training bayesian filters.

I been getting that too, weird shit

hahahahhaha:1orglaugh :BangBang:

yeah maybe just seeing what filters they ca or can not get past :eek7

... you think I should warn my neighbor ??

your email was just verifyed. thanks.

spam assasin and shit adds points for to much html in a message... 80% html is like +2 points... add words to thin it out.. if your mailing from alot of servers into 1 isp.. you don't want nothing to flag the isp as being the same message twice.. so you do random words.. and no computer can ever tell if something is a correct sentance or not.. least i don't think so.. unless the future spam filter will block people that don't know how to spell or form sentances correctly.. :helpme

They are using it to attempt to get past mail filters.

I was reading an discussion about filtering the other day, and they were saying that to get around excessive punctuation (v.1.a.g.r.4) they can start to filter out emails with a number of spelling errors above a statistical norm, but I imagine the processing overhead of that would be pretty intense.

It's definitely possible to determine with a fair approximation whether a piece of text is natural language or not.