Commit

c21203a

TIL why PHP function names are so inconsistent (strlen vs str_replace). In 1994, Rasmus Lerdorf used strlen() as his actual hash function. He added or removed underscores just to balance the hash buckets by character count.
As I understand it, they weren't completely unique, just evenly distributed. With hundreds of functions, unique lengths wouldn't have worked anyway. From what I read, the goal was just to balance the hash buckets so none of them got overloaded, keeping lookups fast. So if a specific length was getting too crowded, he'd tweak the name to shift it into a less populated bucket.

commit c21203a