Cover fe ature 47 january 2012 Published by the ieee computer Society 0018-9162/12/$31. 00 2012 ieee

void drag_item_to(Vector& v,

                 Vector::iterator source,

                  Coordinate p)

{

   auto dest = find_if(v.begin(), v.end(), contains(p));

    if (source < dest)

        // from before insertion point:

        rotate(source, source+1, dest); 

 

    else

        // from after insertion point:

        rotate(dest, source, source+1);

}

That is, find the insertion point using the standard library 

algorithm 

find_if

 and then move the element to the in-

sertion point using the standard library algorithm 

rotate

. 

Obvious once you see it! 

The original solution was complicated enough to 

raise worries about correctness, plus it was completely 

special-purpose. What might happen to such hard-to- 

understand code during maintenance? The improved code 

is shorter, simpler, more general, and runs faster. It uses 

only well-documented standard library facilities known to 

experienced C++ developers.

But why move only one element? Some user interfaces 

allow selecting and moving a collection of elements. And 

why use application-specific notions such as 

Vector

 and 

Coordinate

? Dealing with this more general problem turns 

out to be simpler still: 

template

pair

gather(Iter first, Iter last, Iter p, Predicate pred)

// move e for which pred(e) to the insertion point p

{

     return make_pair(

        // from before insertion point:

        stable_partition(first, p, !bind(pred, _1)),

        // from after insertion point:

        stable_partition(p, last, bind(pred, _1)) 

      );

}

 

The 

Predicate

 determines which elements are moved. 

Admittedly, 

stable_partition

 tends to be used only 

by specialists, but its meaning isn’t obscure: 

stable_