How can I filter out profanity in base36 IDs?

Question

I want to use base36 in a web application I am developing... but as the id is visible to users as a url, I want to filter out profanity. Has anyone solved this? Or is this even a real problem?

Does it make sense just to skip numbers in my database sequence?

Answer 1

Well, rather than try to amass all the swear words possible, just filter out the vowels. That'll leave you plenty of permutations in the space. Admittedly, you've just cut down from base 36 to base 31, but base 31 numbers are valid base 36 numbers assuming the same symbol set (a-z0-9). IF that bothers you, replace the five vowels with some other non-magic 7-bit ascii like !,@,$,% and (.

Granted, you may end up with sh1t and fck, but the profanity is in the mind of the reader.

Answer 2

Why not just use a full-on randomly generated GUID in hexadecimal? No matter what programming language you're working in, this should be easy to generate. And being represented in hexadecimal, I would imagine the chances of generating something that upsets the easily offendable approach zero.