Don't use unsafe ~ character.

[zharikovpro]

TLDR: follow the RFC for URLs. Complete reasoning.

[zharikovpro]

~ can be safely replaced with _

[ai]

@zharikovpro we already have _ in the alphabet.

This article has a problem, it tells that - is safe, but it is not dylang/shortid#96
This is a reason why ~ was added to Nano ID, because it is safe in practice and - is safe in theory.

Do you know any practical problems with ~?

[zharikovpro]

@ai I personally have no real troubles with - or ~. It's all about quote from RFC:

Other characters are unsafe because
gateways and other transport agents are known to sometimes modify
such characters. These characters are "{", "}", "|", "", "^", "~",
"[", "]", and "`".

It may be that ~ also has some weird issues, like one that you referenced for the -. I posted this issue just in case.

Also, I was choosing a short id lib and came to this warning, so this issue may affect library adoption.

[ai]

OK, I will think about it later. Current symbols are safe and work. But I understand that following RFC is good idea.

If you don't like ~ you can replace alphabet. I recommend to not use - as well.

[shesek]

How about replacing ~ in the alphabet with a .?

[shesek]

Though, it appears like RFC 3986 says that ~ is URL safe?

2.3.  Unreserved Characters

   Characters that are allowed in a URI but do not have a reserved
   purpose are called unreserved.  These include uppercase and lowercase
   letters, decimal digits, hyphen, period, underscore, and tilde.

      unreserved  = ALPHA / DIGIT / "-" / "." / "_" / "~"

https://www.ietf.org/rfc/rfc3986.txt

[ai]

I afraid . in the end of the string will be removed from link on Android and etc

[ai]

Also ~ was very popular in 90-2000s as part of user’s directory on the domain: domain.com/~user. I think ~ is safe.