HashSet is implemented by delegating to an internal HashMap,
which is suboptimal, especially from the point of view of "Bloat".
As expressed well in

The Causes of Bloat, the Limits of Health

http://domino.research.ibm.com/comm/research_people.nsf/pages/sevitsky.pubs.html/$FILE/oopsla2007-bloat.pdf

"The standard Java HashSet, which stores
a set of unique values, is implemented by delegating to the
more general HashMap, which in turn points to an array
of HashMap$Entry objects that point to keys and values,
and to each other.7 The delegation of HashSet to HashMap
raises the fixed collection overhead for each of these nested
HashMaps; the overgenerality of HashMap raises both the
fixed and variable overhead costs: the HashMap has book-
keeping fields that are not relevant to this use, and all of the
inner HashMap$Entry objects point to the same (unused)
singleton value."