I'm amazed how quickly spambots find a new site. Especially the comment spammers. When I had this on a temp URL, it had been found within two weeks. My solution? Turn off anonymous commenting. Easy, because it didn't matter, and there were no users to speak of.

But six hours after I pushed this site public, the comment spammers were at it again. Formerly, the site was running on wordpress. There, I used spam karma 2 to stem the tide: it worked really well.

Podcasting is an even worse waste of time than vblogging is the podcast. Imagine a vblog, but without the ability to fast-forward and rewind with visual clues, where two thirds of the content is devoted to useless yammering (one third at the start, and one third at the end). Through in some (lots of!) grainy audio, and you have a podcast.

In the past few months it's been an increasing trend to have sites ask you for your username and password from some service. I've lamented about this in the past regarding youtube and embedding videos, and it's shoved in your face on facebook.

Though, admittedly, youtube has to store your password for some reason, while apparently facebook only uses it then discards it.

Still, I couldn't quite put my finger on why it was a bad idea. Apparently it's in every social networking site, too.

I see lots of sites trying out this new 'Intellitext' service.... tell me it isn't annoying, what with tons of green links everywhere that create huge popups every time you happen to move your mouse over them.

Recently, AOL released a list of its most frequently-seen spam subject lines. While I might not get 556 billion spam messages a year, but my spam is up also (1478 last year, this year: 3351).

So I did my own little investigation on what words appear most frequently in spam email subject lines. A quick grep from the subject lines of my junk folder (in maildir format) and a run through a tokenizer and uniq revealed the following list:

    304 re
    260 you
    239 your
    212 for
    209 a
    182 the
    154 iso
    151 free
    147 is
    144 to
    134 new
    131 confirmation
    114 or
    113 of
    112 st
    109 th
    106 update
    106 this
    104 card
    104 b
     99 rernst
     99 gift
     86 in
     84 q
     84 customer
     83 january
     81 fw
     76 ck
     76 and
     74 stock
     74 get
     74 c
     72 on
     68 shares
     67 r
     63 starbucks
     61 pain
     61 do
     60 at
     54 walmart
     54 here
     52 software
     49 prescripiton

I didn't bother stripping out the two and three character words. Down the list are many many common misspellings of words - these are probably from spammers' attempts to get through spam filters. For interest's sake, the complete list is available here.

I find it interesting that my username 'rernst' appears in quite a few subject lines: very few humans or even automated mail would feature this.... I don't recall seeing it frequently.