ObjectPool.as

package 
{

	public final class ObjectPool
	{
		private var MAX_VALUE:uint; 
       		private var GROWTH_VALUE:uint; 
		private var TYPE:Class;
		private var counter:uint;
		private var pool:Array; 
		
		public function ObjectPool(type:Class, maxPoolSize:uint, growthValue:uint)
		{
			MAX_VALUE = maxPoolSize; 
           		GROWTH_VALUE = growthValue; 
			TYPE = type;
			counter = maxPoolSize;
			
			var i:uint = maxPoolSize;
			
			pool = new Array();
			while (--i > -1)
				pool[i] = new type();
		}
		
		public function getObject():*
		{	
			if (counter > 0) return pool[--counter];
				
			var i:uint = GROWTH_VALUE; 
            		while( --i > -1 ) 
                   		pool.unshift ( new TYPE() ); 
            		counter = GROWTH_VALUE; 
		        return getObject(); 				
		}
		
		public function returnObject(disposedObject:*):void
		{
			pool[counter++] = disposedObject;
		}
		
		public function numberOfObjects():uint
		{
			return counter;
		}		
	}	
}

I can see how this one is better, although more complex than mine.
Here you remove the object from the pool, and its the "user"'s responsibility to return it to the pool later.
My simpler script may break if the user requests more objects than the pool holds.