There’s nothing wrong with having non-normalized representations, that’s why there is equals().
For example, you might have a value class for representing (limited-precision) fractions using two longs internally, for the numerator and denominator. For efficiency trade-off reasons, you don’t want to always shorten the fraction. But now client code can distinguish 2/3 from 4/6 using ==.
Scenarios of that sort are conceivable where this actually leaks sensitive information. In any case, it creates dependencies on implementation details where you don’t want to have them.
When designing a value class, you are now in the dilemma of either always having to normalize the representation, costing performance, or having your class be a funnel for leaking implementation details.
Java can also distinguish a 2/3 object from a 4/6 object using == when they are not value types. It can even distinguish a 2/3 object from a different 2/3 object.
The difference, as I tried to explain, is that identity comparison does not expose the internal representation. So while you can know that two objects with the same value are different objects, you can't know whether they internally represent the value as 2/3 or 4/6. With value classes, however, the latter can expose information about the provenance of a value.
Well. I'd be upset if custom operator==() for plain-old-data structs was removed from C++, but Java never had it to begin with, so for Java, it just means that you have to fall back to using traditional classes (or compare using something other than ==) if you need such "fancy" features.
For example, you might have a value class for representing (limited-precision) fractions using two longs internally, for the numerator and denominator. For efficiency trade-off reasons, you don’t want to always shorten the fraction. But now client code can distinguish 2/3 from 4/6 using ==.
Scenarios of that sort are conceivable where this actually leaks sensitive information. In any case, it creates dependencies on implementation details where you don’t want to have them.
When designing a value class, you are now in the dilemma of either always having to normalize the representation, costing performance, or having your class be a funnel for leaking implementation details.