An equivalence forDoubles which is reflexive (treats allNaNs as equivalent), and treats-0.0 and0.0 as not equivalent; it behaves the same asjava.lang.Double.compare.

Because the behaviour ofDoubles specified by IEEE is not consistent with behaviors required of an equivalence relation forNaN (it is not reflexive), there are two equivalences defined forDouble:StrictEquiv, which is reflexive, andIeeeEquiv, which is consistent with IEEE spec and floating point operations defined inscala.math.

This equivalence may be preferable for collections.