Hacker News new | past | comments | ask | show | jobs | submit login

I don't think it is as easy as you think it is. Consider for example the Scunthorpe Problem [1].

Also consider that people on this site are hackers and so naturally we can all speek Leet [2] ;-) In these days of unicode it's extremely difficult to prevent people posting words that look like swear words, but which on inspection are just unusual characters.

[1] http://en.wikipedia.org/wiki/Scunthorpe_problem

[2] http://en.wikipedia.org/wiki/Leet




arctangent, I was suggesting a lazy filter to set the tone. If you have to open the character map to put the 𝛍 in f__k that gives you time to consider whether it's worth violating the spirit of the forum.

I shortened my comment above but my full suggestion is to apply time-delayed negative karma to posts with Carlin swears rather than ban them. Who is going to spend 45 seconds finding the 𝛍 to avoid −2 karma?


My first point still stands, though. You can't apply a naive filter because people will end up getting penalised for using words like "scrape", "interstitial", "swanky", "spool", etc. ad infinitum.


Word boundary character.


You're very persistent but I'm afraid you're still wrong :-)

Firstly, I gave unicode as a simple example of the lengths that people could go to in order to swear. But it's actually even easier than that in practice. I you want to ban the word X then I can replace it with a slight misspelling which carries the same intent - and when you ban that a naughty user will just pick another one.

(Bad words aren't the problem we should address, it's the intention behind words which is the real problem.)

Secondly, the Scunthorpe problem is more severe than you realise and can't be solved by considering only whole words. For example, some words have more than one meaning: "tit", "gay", "sod", "knob", "cock", "prick", etc. Some or all of those words would be filtered out by a naive filter.

There's also the additional problem that people may not necessarily agree on whether a word is a swear word or not (even if it does not have multiple meanings like the examples I listed below).

For example, in 1977 the Sex Pistols won a court case about the title of their album "Never Mind the Bollocks" in which it was ruled that "bollocks" is not obscene in the UK (but may of course still be considered obscene in other countries!): http://en.wikipedia.org/wiki/Bollocks#Obscenity_court_ruling

And finally, some words mean different things in different coutries. For example, the word "fanny" means one thing in the US and quite another in the UK.

Naive filtering of swear words is never going to work.


never going to work

As I outlined above it's just about the incentive and setting the tone. If you have to write "siht" you'll think twice about whether it needs to be said.

I might do well to add that I am not suggesting this for prudish reasons. I took a creative writing class in college and the teacher banned curses and drugs. Intro to Creative Writing students otherwise substitute force for substance.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: