…and by "l10O359" you mean "1lO0359"

In interacting with a domain registrar this morning I received a message that looked approximately like:

A few things come to mind:

1. This code is 27 characters long. The longer the code the more chance you have of someone improperly copying it, or not being able to read it over the phone. This applies both to the customer and to the support staff. This code has 62^27 combinations, which seems like overkill. An eight character code using only uppercase letters would be sufficient for 208 billion combinations (26^8), and is way more usable to us humans.

2. This code mixes uppercase and lowercase. Uppercase is often more readable, and support staff can always assume that when a customer says “L” they mean uppercase. You can also build the systems to ignore case, which helps even more.

3. This code mixes numbers and letters. Normally this isn’t a problem but ‘1’ and ‘l’ (lowercase ‘L’) look alike in many fonts. Zeros and ‘O’ also get confused in some fonts. It is a bad idea to assume anything about the environment a customer has, including fonts.

So in short, keep it short and it’ll be easier to deal with.