This is one of those pages where I will not go into the full history of enum. It started with C, but so did most of the C++ language. In its original form, it was a vestigal hangover from C -- little more than a list of names for integers. It seems like the ones from textbooks looked like this:
enum { zero = 0, one, two, ten = 10 };
Most of the ones in code did, too, and it raised the question "What is the advantage over #define?
#define zero (0) #define one (1) #define two (2) #define ten (10)
The underlying truth of this presentation of enum has a lot to do with failure to adopt. The underlying truth being that enum was little more than an integer tarted up for the type-safe, object oriented zealots. The veneer of type safety could be eroded by the simplest of statements:
if (x > two && x < ten) ...
The new world order of C++11 has arrived.
If you want type safety, try the class keyword in the enum context:
enum class X { ... one, two ... };
..
X myEnumObj;
The compiler will never accept implicit conversions to integers of your objects of this type. Additionally, accidental (or intentional) use of the scope names for the values the enum may take will not be allowed to migrate their way into inappropriate situations. You will get a compilation error.
Although the above is the big change, there are a couple of others. enum is now a name that provides a scope of its own, so the names that appear inside the braces above are named X::one and X::two preventing collisions between common names (that always seem to be the ones we choose).
Additionally, it is possible to specify the underlying variety of "int" that is being used. This can be done by this syntax of using the colon after the type name.
enum class X : unsigned long long { ... };