What is the reason for BitSet's size() method?

Question

Is there a use case for the size() method on the java.util.BitSet class?

I mean - the JavaDoc clearly says it\'s implementation dependant, it returns the size of the int

Answer 1

If the size method wasn't designed by Java creators as public, it would still undoubtedly exist as a private method/field. So we are discussing its accessibility and maybe naming.

Java 1.0 took a lot of inspiration, not just the procedural syntax, from C/C++. In the C++ standard library, the counterparts to BitSet's length and size also exist. They are called there size and capacity, respectively. There is rarely any hard reason to use capacity in C++, and even less so in a garbage collected language such as Java, but having the method accessible is still arguably useful. I will explain in Java terms.

Tell me, what is the maximum number of machine instructions ever needed for executing a BitSet operation such as set? One would like to answer "just a handful", but this is only true if that particular operation does not result in reallocation of the whole underlying array. Theoretically, the reallocations turn a constant time algorithm into a linear time one.

Does this theoretical difference have much practical impact? Rarely. The array usually doesn't grow too often. However, whenever you have an algorithm operating over a gradually growing BitSet with an approximately known final size, you will save on reallocations if you pass the final size already to the BitSet's constructor. In some very special circumstances this may even have a noticeable effect, in most circumstances it does not hurt.

• set then has constant time complexity - calling it cannot ever block the application for too long.
• if just one extremely large BitSet instance is using up all your available memory (by design), swapping may start noticeably later dependending on how your JVM implements the growth operation (with or without an extra copy).

Now imagine that you operate on many BitSets, all of which have been allocated with a target size. You are constructing one BitSet instance from another and you want the new one share the old one's target size as you know you will be using them side by side. Having the size method public makes this easier to implement cleanly.