Get the text nodes out of a document, ignoring the ones that are in Elements where the text value aren't likely to be valuable (like <script> tags) and nodes containing just whitespace.

text_nodes.js

function getLegitTextNodes() {
  if (!document.createTreeWalker) return [];
  
  var blacklist = ['SCRIPT', 'OPTION', 'TEXTAREA'],
      textNodes = [],
      walker = document.createTreeWalker(
        document.body,
        NodeFilter.SHOW_TEXT,
        function excludeBlacklistedNodes(node) {
          if (blacklist.indexOf(node.parentElement.nodeName.toUpperCase()) >= 0) return NodeFilter.FILTER_REJECT;
          if (String.prototype.trim && !node.nodeValue.trim().length) return NodeFilter.FILTER_SKIP;
          return NodeFilter.FILTER_ACCEPT;
        },
        false    
      );
    
  while(walker.nextNode()) textNodes.push(walker.currentNode);
  
  return textNodes;
}

// usage:
// getLegitTextNodes().forEach(function (node, i) { console.log(node.nodeValue) })