Cursing

Dec. 20th, 2005 01:38 pm
jack: (Default)
[personal profile] jack
Wow, I use the word 'dick' a lot. (here, if you want to see the meme) This surprised me: afaik I don't use the word 'dick' at all.

It's not something I tend to swear with. I generally use something inventive, or 'fuck'.

It's not something I call people -- to me someone being a 'dick' has a subtle distinction from similar insults, suggesting someone who is obnoxious, and not deliberately maelevolent, but specifically accepts their behaviour and makes no effort to change it.

It's not something I call a penis except in occasional jest. I've annoyed or endeared a couple of people by being basically unable to call a penis anything but a spade penis.

And then I remembered - I had a long post about translating spotted dick into spanish. That was probably it.

It reminded me of Gene Wiengarten testing his spam filter. He found several emails about female dogs, gut men porn at monmouth, and the like vanishing, and then when he was thinking the threshold was too high, got a "see a teen girl do it with her horse its[1] free", and commented "They still have some kinks to work out of the system. As it were."

This sort of thing keyword filtering isn't going to catch, and indeed would be quite difficult to filter for even with fairly decent language processing, though normally gives itself away by having html, pictures, non-existant words, links, or 1En copies.

[1] sic.

Date: 2005-12-20 06:53 pm (UTC)
From: [identity profile] feanelwa.livejournal.com
You could get rid of that one by filtering out the phrase "see a teen girl/teen girls do it with". There don't seem to be a lot of other uses for that phrase.

Date: 2005-12-21 02:48 am (UTC)
From: [identity profile] captain-aj.livejournal.com
At the very least, that kind of thing would filter out anybody talking *about* spam ... like if this post were in e-mail form :-). Or people with a disturbing sense of humour. Which is why modern spam solutions tend to learn ham as well as spam, so in that instance they could say "keywords point to it being spam, but there's so much pre-amble that looks like things we've seen (only) in legitimate e-mail before, so on the balance of probabilities it's genuine".

Date: 2005-12-21 01:57 pm (UTC)
From: [identity profile] cartesiandaemon.livejournal.com
Indeed.

And there's so many different phrasings. Even If you could ban "see a teen girl do it with" they can say "watch as lucy and dobbin get jiggy with it" which is completely different on any low level, you can't do everything.