Obtain an XML element by index rather than Tag

I'm creating a video library and have an XML document split up by video category such as:

<video>
    <comedy>
        <url>bla</url>
        <title>blabla</title>
    </comedy>

    <action>
        <url>bla</url>
        <title>blabla</title>
    </action>
</video>

And so on. I use an XMLHttpRequest to getElementsByTagName() for the genre I want and it is working fine.

My question is: I want to create a 'Most Recent' category, that would just pick the first 16 (or however many) off the top of the XML file, regardless of category. Is there a way to accomplish this?

Iterate over the childNodes of the root-element, and check if the node is a element-node(childNodes will also return textNodes for the whitespaces)

var doc=xml,//the xml-document
            childs=doc.documentElement.childNodes,
               i=-1,
               j=16;//how many nodes you like to get

    while(j && childs[++i])
    {
      if(childs[i].nodeType===1)//check if we got a element-node
      {
        j--;
        //do something with the element
        console.log(childs[i].tagName);
      }
    }

In some browsers there may also be a children-property which only returns children that are element-nodes, but this is no standard so I wouldn't suggest to use it.

When you use a library like jQuery it would be much easier, to have the same result you only need:

$('>*:lt(16)',xml.documentElement)
  .each(function(i,o){console.log(o.tagName);});