Javascript word-count for any given DOM element

Paolo Bergantino's second solution is incorrect for empty strings or strings that begin or end with whitespaces. Here's the fix:

var count = !s ? 0 : (s.split(/^\s+$/).length === 2 ? 0 : 2 +
    s.split(/\s+/).length - s.split(/^\s+/).length - s.split(/\s+$/).length);

Explanation: If the string is empty, there are zero words; If the string has only whitespaces, there are zero words; Else, count the number of whitespace groups without the ones from the beginning and the end of the string.

If you know that the DIV is only going to have text in it, you can KISS:

var count = document.getElementById('content').innerHTML.split(' ').length;

If the div can have HTML tags in it, you're going to have to traverse its children looking for text nodes:

function get_text(el) {
    ret = "";
    var length = el.childNodes.length;
    for(var i = 0; i < length; i++) {
        var node = el.childNodes[i];
        if(node.nodeType != 8) {
            ret += node.nodeType != 1 ? node.nodeValue : get_text(node);
        }
    }
    return ret;
}
var words = get_text(document.getElementById('content'));
var count = words.split(' ').length;

This is the same logic that the jQuery library uses to achieve the effect of its text() function. jQuery is a pretty awesome library that in this case is not necessary. However, if you find yourself doing a lot of DOM manipulation or AJAX then you might want to check it out.

EDIT:

As noted by Gumbo in the comments, the way we are splitting the strings above would count two consecutive spaces as a word. If you expect that sort of thing (and even if you don't) it's probably best to avoid it by splitting on a regular expression instead of on a simple space character. Keeping that in mind, instead of doing the above split, you should do something like this:

var count = words.split(/\s+/).length;

The only difference being on what we're passing to the split function.