Initializing Arrays


Initializing large arrays of data before use is always cumbersome but it seems to be unavoidable.

The first types of solutions fill the array with a value that denotes an empty entry. While this sounds straightforward, the choice of that value is not always easy. For floating points numbers, zero may or may not be a good candidate. If convenient (as a zero floating point value is encoded as binary zeroes filling the variable) it may be difficult in some contexts to use because zero may be valid. Fortunately, there’s always the possibility to initialize the array with NaNs, which can be tested and used correctly (but you need the POSIX functions to do that).

Read the rest of this entry »