Question

Possible Duplicate:
Why does the C++ standard algorithm “count” return a ptrdiff_t instead of size_t?

There is algorithm std::count/std::count_if in standard C++.

template<class InputIterator, class T>
typename iterator_traits<InputIterator>::difference_type
count(InputIterator first, InputIterator last, const T& value);

template<class InputIterator, class Predicate>
typename iterator_traits<InputIterator>::difference_type
count_if(InputIterator first, InputIterator last, Predicate pred);

Effects: Returns the number of iterators i in the range [first,last) for which the following corresponding
conditions hold: *i == value, pred(*i) != false.

difference_type is iterator's difference_type, which can be negative, but count can return only value >= 0. Why difference_type and not size_t for example?

Answer 1

In theory you may have a tremendous sequence whose number of elements can only be represented with 128 bits. Assuming the implementation supports a corresponding integer type, it is quite likely that size_t use a 64 bit type. However, the iterator for this huge sequence could use a 128 bit integer. Note, that the sequence isn’t necessary represented in the memory of any individual computer. It may be split across multiple huge databases.

You might also have a relatively small computer supporting only a 32 bit type with reasonable performance. For full standard conformance it might have to support a 64 bit type but it can be desirable to rather support faster computations using a representation natively supported by the platform. That is, size_t may not be the optimal choice. When creating iterators you generally know what size needs to be supported.