* Quadrupling the size improves average dictionary sparseness
* (reducing collisions) at the cost of some memory and iteration
* speed (which loops over every possible entry). It also halves
-| * the number of expensive resize operations in a growing dictionary.
+ * the number of expensive resize operations in a growing dictionary.
*
* Very large dictionaries (over 50K items) use doubling instead.
* This may help applications with severe memory constraints.