News for hackers. Stuff that matters.
Re:Let’s talk about usernames by grzm on Sunday February 11, 2018 @11:49PM (#16356612)
The article does a very intelligent job of disentangling system identifier, login, and display names (and helpfully links to a discussion of the tripartite identity pattern†), which obviates a lot of what you describe here.

† http://habitatchronicles.com/2008/10/the-tripartite-identity...

Re:Let’s talk about usernames by saurik on Sunday February 11, 2018 @11:56PM (#16356638)
Yes, but it does not bother to motivate why people should use "tripartite identity" (and even that linked article doesn't really try to make an argument); and then it goes on to talk about why uniqueness is hard from a technical level, leading people to be responding with comments like "I just included a library in my stack or chose the right type for my PostgreSQL column that fully handled the case and Unicode mapping, so this wasn't a big deal".

The real problem is that unique and permanent usernames serve as tatoos (which people later may find to be humiliating or depressing), disadvantage late-comer non-technical users (who will almost never have a good username and almost never will have the same username on two websites), and lead to weird problems with assumptions people make about what usernames even mean (that they are a signal for identity) that are simply not true.

Re:Let’s talk about usernames by Natanael_L on Monday February 12, 2018 @09:27AM (#16358734)
How would you handle making them non-permanent, though?

Expiring hotmail addresses have been problematic in many cases, and any unique lookup string will eventually be stored somewhere by someone and assumed still valid later on.

It's not even solved in full for phone numbers, despite everybody knowing they can expire and be reassigned - since long before our own lifetimes.