From Forum Nokia Wiki

Problem Statement : If your Widget is connecting to a remote HTML-site through XMLHttpRequest(), the function responseXML will not work, as the response is not pure XML. Thus the following example is showing a smooth solution on how to deal with this problem and parse the HTML as it was an XML tree.

The code shows a basic Widget, that connects to a remote (or a local file for debug reasons) HTML-Site and parses the content. As the returned HTML-Code is not XML-well formed the 'responseText()' method is used to retrieve the plain HTML-Code as text, which then is assigned to a dynamically created div using 'document.createElement('div');'

You can download the full Widget here.

To connect to a remote site, you can use this code snippet:

// defined global, as required in call-back as well.
var xmlHttp = null;
 
function openConnection()
{
	try{
 
		// create a new connection object
		 xmlHttp = new XMLHttpRequest();
 
		// reference to Call-Back Function;
		xmlHttp.onreadystatechange = httpCallBack;
 
		// url you connect to; can also be a local file
		var url_frame = "local.txt";
 
		// open the URL now
		xmlHttp.open('GET', url_frame, true);
 
		// and send GET request
		xmlHttp.send(null);
 
	}
	catch(e)
	{
 
	}
}

Asynchronously (that's the first 'A' in AJAX ;-) the Call-Back function 'httpCallBack' of the Widget is called

// Call-Back Function for the XMLHttpRequest() Request
function httpCallBack()
{
 
	if(xmlHttp.readyState != 4)
	{
		// tell the user, that the your widget is still
		// waiting for the response to be completed
	}
 
 
	// readyStatus=4 Means completed
	if (xmlHttp.readyState == 4) {
 
 
		// get HTTP Data	
		try
		{
			// first check, if there is a response at all
			var resultXml = xmlHttp.responseText;
			if(!resultXml)
				// put some error message here. "repsonseText Failed or alike"; 
 
			else
			{
 
				// now create a new div (actually a node!) dynamically
				// and do some assignments of parameters
				var newdiv = document.createElement('div');
				newdiv.setAttribute('id','temp');
 
				// put the response of the Site into this div
				// this allows us to treat the result as a node and therefore
				// use eg 'getElementsByTagName' on the HTML, that is not XML
				// conform! That's a very nice thing to do for that purpose.
				newdiv.innerHTML =  xmlHttp.responseText;
 
				// now you can parse your content as you like using the HTML Code
				// as XML structure
				// eg: newdiv.getElementsByTagName("td"); 
 
			}// end-else
 
		} // end-try
  	catch (e) {
	    // do some exception handling here. 
    }
 
 
	}
 
}