Safe Characters all language about hashids HOT 1 CLOSED

Yeah I've thought about that too. So, currently Hashids does not eliminate, but uses these as separators: "c, s, f, h, u, i, t". That way if you want to encode [1, 2, 3], and 1 encodes to "A", 2 encodes to "B", and 3 encodes to "C", your hash might be: "AsBuC".

If I were to remove all vowels, calculation would be faster (which is a plus), but the generated hashes would be longer - additionally they might not look as "random" since there'd be less characters (and some people might not even like that their hashes don't have vowels).

But if I were to use all vowels as separators (instead of the current characters), that doesn't help with the curse words either -- if 1 encodes to "F" and 2 encodes to "CK", we'd have a problem when "u" is the separator.

So, I'm hesitant to use the approach of eliminating all vowels because 1) I'm not sure the solution outweighs the problem and 2) I haven't noticed it being that big of a problem yet (I've been using Hashids at 3 companies so far - with tons of generated hashes and have yet to see a bad word - but, on the other hand I've only been keeping an eye out for English words).

But, ultimately it all depends on the community. If there's a strong support for it (whether it's this approach or any other), I'll gladly readjust Hashids. And if you decide to modify Hashids to use this technique, tell me so I can link to your repo so people have some alternatives.

from hashids.