All the Perl that's Practical to Extract and Report

Markov Spammer

VSarkiss on 2003-12-02T00:50:03

I don't know why I find it amusing when I see the lengths spammers will go to, but I do. Today one spam email made it through my filters (for pony rides, of all things), and the body contained one paragraph, which I reproduce verbatim:

baklava desolater ammoniac reginald dissociable crabapple argon australia caliber distant spoonful canto coachwork exile leftover compline comparative basophilic onondaga asuncion vivo transferable involution crevice crack loan czar futile horn bowmen musicology mold blockade stigmata chocolate enable greet yeasty pharmacist mess
It looks like they used multi-lingual dictionaries; "asuncion" is Spanish, isn't it? And I have no clue what "onondaga" means....

Geography

vsergu on 2003-12-02T01:02:15

Asunción is a Spanish word, but it's also a placename (the capital of Paraguay), as is Onondaga (city and county in New York, as well as an Indian tribe). I think the dictionary is just English.

The spam isn't very Markovian. There are hardly any pairs of words that are likely to occur together in real text, much less longer chains.

Re:Geography

hex on 2003-12-02T12:56:58

At some unspecified point in the future I will upgrade Spam Trap to use Markov chains, because they are groovy.