Although I'm aware that display names are not unique, this is the first time I've come across a question answered by a user with the same display name as the OP (here).
The problem this poses is that if I want leave a comment on the answer that notifies the asker, I have to play a guessing game:
There's no clear way for me to tell which Kareem is the real Kareem (short of a certain arcane incantation) and I wouldn't want to mistakenly shoot (an answer to) the wrong one.
This is confusing (if infrequent). I can't tell whether it actually breaks notifications but even if a comment using @Kareem would "correctly" notify both users, a system that allows arbitrary identifiers should be prepared to handle collisions a bit more gracefully. How many people read the @replies FAQ? (More importantly, how many remember all of the implementation details?)
What I'd like to see is a slightly expanded notification syntax; for example, @: could allow notification by the user ID such that I would use @:2561 to notify the OP in the above conversation.
Being able to type @: to show autocomplete options as Kareem:2561 and Kareem:1661 instead of just Kareem and Kareem would be a nice additional functionality.
I'm assuming here that an arbitrary display name could be 2561 or user2561 but could not start with a symbol. Another approach that avoids expanding the syntax would be to reserve display names of the format ^user\d+$ to avoid spoofing the "default" display name of another user and then allow @-notification by either of the current and default display names.
One additional benefit of this feature is that it would be much easier to distinguish when a reply is made to a deleted comment vs. a comment from a user whose display name had later changed. Again, this will be infrequent, and it relies on people using the feature - but if user2561 changes their display name on realizing there's an active user already going by "Kareem," the @reply by the other "Kareem" will start to look a bit schizophrenic. It's in just those cases where a user is most likely to want the alternative syntax that it might be expected to carry this added benefit.
