Low-Level Programming

The “low level” in low-level programming refers to the level of abstraction, not to the quality of the programming. Low level means closer to the bits and bytes of computer hardware and machine language. Low-level programming is important for embedded programming and for increasing the efficiency of some operations. C++11 offers some aids to those who do low-level programming.

One change is relaxing the constraints on what qualifies as “Plain Old Data,” or POD. In C++98, a POD is a scalar type (a one-value type, such as int or double) or an old-fashioned structure with no constructors, base classes, private data, virtual functions, and so on. The idea was that a POD is something for which it’s safe to make a byte-by-byte copy. That is still the idea, but C++11 recognizes one can remove some the C++98 restrictions and still have a viable POD. This helps low-level programming because some low-level operations, such as using the C functions for byte-wise copying or binary I/O, required PODs.

Another change is making unions more flexible by allowing them to have members that have constructors and destructors, but with some restrictions on other properties, for example, not allowing virtual functions. Unions are often utilized when minimizing the amount of memory used is important, and the new rules allow programmers to bring more flexibility and capability to these situations.

C++11 addresses memory alignment issues. Computer systems can restrict how data is stored in memory. For example, one system might require that a double value be stored at an even-numbered memory location, whereas another might require the storage to begin at a location that is a multiple of eight. The alignof() operator (see Appendix E, “Other Operators”) provides information on the alignment requirements for a type or object. The alignas specifier provides some control over the alignment used.

The constexpr mechanism expands the ability of the compiler to evaluate during compile time expressions that evaluate to a constant value. The low-level aspect of this is to allow const variables to be stored in read-only memory, which can be particularly useful in embedded programming. (Variables, const or otherwise, that are initialized during runtime, are stored in random-access memory.)