It's certainly an interesting data set, though it has no concept of severity. As far as I can tell, "doodoo" is the same as some racial slurs: we're 100% certain they're bad words.

Could have been in a language agnostic format (eg. csv)

Typical cuss filter UX:

types something in live chat

some random word from the sentence gets censored out

"Why did this just got censored out?"

check urban disctionary

"Why?????"

Bonus points if its regular ethnonyms that are classified as profanities, so people from that place are having big trouble to tell where they are from.

Was really amused to see that a paper had English's most prominent profane word in it's abstract on arXiv last month for the first time:

https://arxiv.org/search/?query=fuck&searchtype=all&source=h...

though somebody did slip in a use in a comment earlier.

Something we have had to deal with in managing educational software with a writing aspect is trying to manage what is offensive to who, in what context and where is not universal at all.

One of the most prime examples, at one point a number of terms related to homosexuality had made it onto the list at the request of a larger district. These are also terms that are being reclaimed, and it was... a difficult problem to try to satisfy everyone, and it did upset other districts. I believe their patterns were all but removed eventually.

We have a fought over the list of definitions and every change provoked controversy. Our current solution is just that we mark items for teacher review but don't tell them why. We don't say they are offensive, we don't say what the problematic words are. We just say it might need review. That's worked pretty well so far.

All this is to say, policing speech is a problem best avoided.

Good to know that "This package is safe."

When it comes to security, the only thing that beats warm fuzzy words are shiny security seals.

I legit thought this said "... rating of success" meaning how likely the project was to be successful on some metric based on the profane words therein. I recall there was a study(?) akin to that for the Linux kernel, as a frame of reference

Never realized the "chunky" in my chunky peanut butter was so profane. /s

Where does the rating come from? Do you understand what all those words mean? It looks like you copied someone's rather subjective opinion. Because e.g. "bollo" and "caliente" aren't inherently profane in Spanish. Or do people think the hot water tap is leering at them? "Oye, tia, que caliente qu'et-ta!"

“Sureness” is a not really a word, I had to read through to understand what they meant. “Certainty” or “confidence” would be clearer.

Based on a list of (in part) profane words which includes:

addict africa amateur american angry arab

"Beaver" unlikely to be used in profanity, eh?

It seems to require specifying all spelling variants of a word https://github.com/words/cuss/blob/6bab3fef250481e34ba55bc40...

And then fails to do that for words that are not uncommonly written with a space https://github.com/words/cuss/blob/6bab3fef250481e34ba55bc40...

Making this a complete list will probably be a challenge when it needs to be a byte-for-byte match

Somewhat related: What is with the rampant cursing nowadays? In the US people are openly saying f-word in professional settings, in public to strangers or acquaintances, in writing and video... seemingly everywhere even in calm normal conversations.

I don't remember it being like this decades ago. Is it just me? I remember people used to curse only in private conversation, when angry, and never at the office in meetings and professional contexts.

I'm confused as to the purpose of all the zeros. Since this is far, far from a complete list of all English words, what's the difference between a word not being on the list vs a word being a zero?

I can kind of see "was this a word they considered and scored, vs. not considered?" when trying to assess whether the project is comprehensive, but from a programming standpoint, it just seems like it's going to have a lot of useless overhead, since by the time I'm looking up the word I don't care whether it's a zero or a miss.

(I also find the scoring of "2" for many of the words to be weird, like "yank," "chug," "looser" etc. as they can all have perfectly normal meanings.)

Helpful tool for car makers.

Would have probably saved them from the Mitsibishi Pajero, Ford Pinto, Mazda Laputa

Downside is, it doesn’t analyze phonetics afaict. The hebrew Volkswagen Beetle (Hipushit) would have passed as fine.

The Dutch word 'kunt' (je kunt = you can) gets censored in WoW because of 'cunt'. That is, if you have mature language filter on. I have this on because I have no interest in raging kids in said game, but I do want to read simple, common Dutch words. Annoys me to this day. CS gave the obvious answer (WONTFIX, with obvious workaround disabling the mature language filter altogether). It could be solved easily by looking at context instead of simple blacklisting. I connect from a Dutch IPv4. I sometimes talk Dutch. The same would be true for the other endpoint.

I am reminded of the late great Eudora, a Mac mail program. Late versions would flag ‘offensive’ terms in both outgoing and incoming messages. A hidden option setting would cause it to read aloud all flagged text.

Nit: why is Portuguese named "European Portuguese"? If anything, the language spoken in Brazil should be called "American Portuguese".

These articles always remind me of some code for particle physics simulations. It was full of variables called anal_this and anal_that (because analysis).

Someone put a comment "stop calling variables ANAL!!! This is physics not an orgy!!!!"

I may have a copy on a disquette somewhere :)

Do we know how exactly are these certainty ratings determined?

Edit: seems like it’s all arbitrary? E.g. in a PR[1] I saw random new words get added with no explanation of why a certain rating gets assigned.

[1]: https://github.com/words/cuss/pull/43/files (nsfw too).

I had a look at the French ones.

1/4 is normal everyday cussing

1/4 is cussing when the team is losing, but there are children around

1/4 are from the 17th century and I had a good laugh

1/4 are useful when driving

3 are actually bad (just an estimation :)).

The thing with French is that the cussing is quickly funny.

Looked at french words, most have a rating of 2 (mostly profane) even for words that are not profane at all (ex: envoyer => send), words that have a profane second meaning, but their non-profane meaning is also in common use (ex: morue => cod). Also "retard" just means "delay", I have never heard it used as profanity, maybe in Quebec? ("retard" in English would translate to "attardé" in French)

Do not forget words like "Dragon" or "Figma".

there may need to be additional localization of english. i am in awe of the english and scots' ability to turn simple words into curses ("spanner" might be one of my favorite insults, when spoken by UK folks, but it's insufferable coming from non-UK/GB folks).