Pemrograman Berbasis Web

Semester Gasal TA 2013/2014

Ajax

Basic objects necessary

Setting up the XMLHttpRequest object
Making the call
How the server responds
Using the reply
XML basics

The usual way we operate in the Web

Typical browsing behavior consists of loading a web page, then selecting some
action that we want to do, filling out a form, submitting the information, etc.
We work in this sequential manner, requesting one page at a time, and have to wait
for the server to respond, loading a whole new web page before we continue.
This is also one of the limitations of web pages, where transmitting information
between a client and server generally requires a new page to be loaded.

JavaScript is one way to cut down on (some of) the client-server response time, by
using it to verify form (or other) information before its submitted to a server.

One of the limitations of JavaScript is (or used to be) that there was no way to
communicate directly with a web server.

Another drawback to this usual sequential access method is that there are many
situations where you load a new page that shares lots of the same parts as the old
(consider the case where you have a menu bar on the top or side of the page that
doesnt change from page to page).

Things change

Until recently, we didnt have any alternative to this load/wait/respond method of

web browsing.

Ajax (sometimes written AJAX) is a means of using JavaScript to communicate with

a web server without submitting a form or loading a new page.

Ajax makes use of a built-in object, XMLHttpRequest, to perform this function.

This object is not yet part of the DOM (Document Object Model) standard, but is
supported (in different fashions) by Firefox, Internet Explorer, Safari, Opera, and
other popular browsers.

The term Ajax was coined in 2005, but the XMLHttpRequest object was first
supported by Internet Explorer several years before this.

Ajax

Ajax stands for Asynchronous JavaScript and XML.

The word asynchronous means that the user isnt left waiting for the server the
respond to a request, but can continue using the web page.

The typical method for using Ajax is the following:

1) A JavaScript creates an XMLHttpRequest object, initializes it with
relevant information as necessary, and sends it to the server. The script
(or web page) can continue after sending it to the server.
2) The server responds by sending the contents of a file or the output of a
server side program (written, for example, in PHP).
3) When the response arrives from the server, a JavaScript function is
triggered to act on the data supplied by the server.
4) This JavaScript response function typically refreshes the display using the
DOM, avoiding the requirement to reload or refresh the entire page.

The Back End

The part of the Ajax application that resides on the web server is referred to as the
back end.
This back end could be simply a file that the server passes back to the client, which
is then displayed for the user.
Alternatively, the back end could be a program, written in PHP, Perl, Ruby, Python,
C, or some other language that performs an operation and sends results back to
the client browser.
An XMLHttpRequest object can send information using the GET and POST
methods to the server in the same way that an HTML form sends information.
Recall from our previous discussions that the GET request encodes the information
inside of the URL, while a POST request sends its data separately (and can contain
more information than a GET request can).

Writing an Ajax application

We have to write the front end of the application in JavaScript to initiate the
request.
The back end, as mentioned, processes the request and sends its response back
to the client. The back end is typically a short program we write for performing
some dedicated task. This could be scripted in any language that is capable of
sending back communication to the browser, like PHP or Perl.
We also need to write the JavaScript response function for processing the
response and displaying any results (or alterations to the web page).

The x in Ajax stands for XML, the extensible markup language. XML looks like
HTML, which is no mistake as the latest versions of HTML are built upon XML. The
back end could send data back in XML format and the JavaScript response
function can process it using built-in functions for working with XML. The back end
could also send plain text, HTML, or even data in the JavaScript format.

We will discuss some of these methods for sending data back to the requesting
client and how it can be processed.

The XMLHttpRequest object

The XMLHttpRequest object is the backbone of every Ajax method. Each

application requires the creation of one of these objects. So how do we do it?

As with most things in web programming, this depends upon the web browser that
the client is using because of the different ways in which the object has been
implemented in the browsers.

Firefox, Safari, Opera, and some other browsers can create one of these objects
simply using the new keyword.
<script type="text/javascript">
ajaxRequest = new XMLHttpRequest();
</script>

The XMLHttpRequest object (cont.)

Microsoft Internet Explorer implements this object using its proprietary ActiveX
technology. This requires a different syntax for creating the object (and can also
depend upon the particular version of Internet Explorer being used).

To handle different types of browsers, we use the

try { . . . } catch (error) { . . . }

format. The try section attempts to execute some JavaScipt code. If an error
occurs, the catch section is used to intervene before the error crashes the
JavaScript (either to indicate an error has happened, or to attempt something else).

To create one of these objects we can use a sequence of try. . . catch blocks,
attempting different ways to create an XMLHttpRequest object.

The XMLHttpRequest object (cont.)

function getXMLHttpRequest()
/*
This function attempts to get an Ajax request object by trying
a few different methods for different browsers. */
{
var request, err;
try {
request = new XMLHttpRequest();
// Firefox, Safari, Opera, etc.
}
catch(err) {
try {
// first attempt for Internet Explorer
request = new ActiveXObject("MSXML2.XMLHttp.6.0");
}
catch (err) {
try {
// second attempt for Internet Explorer
request = new ActiveXObject("MSXML2.XMLHttp.3.0");
}
catch (err) {
request = false; // oops, cant create one!
}
}
}
return request;
}

If this function doesnt return false then we were successful in creating an

XMLHttpRequest object.

The XMLHttpRequest object (cont.)

As with any object in JavaScript (and other programming languages), the

XMLHttpRequest object contains various properties and methods.

We list the most important of these properties and methods on the next pages.

The main idea is that the properties are set after the object is created to specify
information to be sent to the server, as well as how to handle the response received
from the server. Some properties will be updated to hold status information about
whether the request finished successfully.

The methods are used to send the request to the server, and to monitor the
progress of the request as it is executed (and to determine if it was completed
successfully).

XMLHttpRequest object properties

Property

readyState

onreadystatechange

responseText

responseXML

status

statusText

Description
An integer from 0. . .4. (0 means the call
is uninitialized, 4 means that the call is
complete.)
Determines the function called when the
objects readyState changes.
Data returned from the server as a text
string (read-only).
Data returned from the server as an XML
document object (read-only).
HTTP status code returned by the server
HTTP status phrase returned by the server

We use the readyState to determine when the request has been completed, and then
check the status to see if it executed without an error. (Well see how to do this
shortly.)

XMLHttpRequest object methods

Method

Description
open('method', 'URL', asyn) Specifies the HTTP method to be used (GET
or POST as a string, the target URL, and
whether or not the request should be
handled asynchronously (asyn should be
true or false, if omitted, true is
assumed).
send(content)
Sends the data for a POST request and
starts the request, if GET is used you
should call send(null).
setRequestHeader('x','y')
Sets a parameter and value pair x=y and
assigns it to the header to be sent with
the request.
getAllResponseHeaders()
Returns all headers as a string.
getResponseHeader(x)
Returns header x as a string.
abort()
Stops the current operation.

The open object method is used to set up the request, and the send method starts the
request by sending it to the server (with data for the server if the POST method is used).

A general skeleton for an Ajax application

<script type="text/javascript">
// ***** include the getXMLHttpRequest function defined before
var ajaxRequest = getXMLHttpRequest();
if (ajaxRequest) {

//

if the object was created successfully

ajaxRequest.onreadystatechange = ajaxResponse;
ajaxRequest.open("GET", "search.php?query=Bob");
ajaxRequest.send(null);
}

function ajaxResponse() //This gets called when the readyState changes.

{
if (ajaxRequest.readyState != 4) // check to see if were done
{ return; }
else {
if (ajaxRequest.status == 200) // check to see if successful
{
// process server data here. . . }
else {
alert("Request failed: " + ajaxRequest.statusText);
}
}
}
</script>

A first example

Heres an example to illustrate the ideas weve mentioned (inspired by an example

in the book Ajax in 10 Minutes by Phil Ballard).
The main idea is that were going to get the time on the server and display it to the
screen (and provide a button for a user to update this time). The point I want to
demonstrate here is how to use Ajax to do this update without updating/refreshing
the entire webpage.
We use a (very) small PHP script to get the date from the server, and return it as a
string as a response to the request. Here is the script:
<?php
echo date('H:i:s');
?>

I saved this as the file telltime.php.

The HTML file and JavaScript code follows.

<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN"

"http://www.w3.org/TR/xhtml11/DTD/xhtml1-strict.dtd">
<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
<head>
<title>Ajax Demonstration</title>
<style>
body {
background-color: #CCCCCC;
text-align: center;
}
.displaybox {
margin: auto;
width: 150px;
background-color: #FFFFFF;
border: 2px solid #000000;
padding: 10px;
font: 1.5em normal verdana, helvetica, arial, sans-serif;
}
</style>
<script type="text/javascript">
var ajaxRequest;
function getXMLHttpRequest()
/*
This function attempts to get an Ajax request object by trying
a few different methods for different browsers. */
{
// same code as before. . .
}

function ajaxResponse() //This gets called when the readyState changes.

{
if (ajaxRequest.readyState != 4) // check to see if we're done
{ return; }
else {
if (ajaxRequest.status == 200) // check to see if successful
{
document.getElementById("showtime").innerHTML =
ajaxRequest.responseText; }
else {
alert("Request failed: " + ajaxRequest.statusText);
}
}
}

function getServerTime()
// The main JavaScript for calling the update.
{
ajaxRequest = getXMLHttpRequest();
if (!ajaxRequest) {
document.getElementById("showtime").innerHTML = "Request error!";
return;
}
var myURL = "telltime.php";
var myRand = parseInt(Math.random()*999999999999999);
myURL = myURL + "?rand=" + myRand;
ajaxRequest.onreadystatechange = ajaxResponse;
ajaxRequest.open("GET", myURL);
ajaxRequest.send(null);
}
</script>
</head>

<body onload="getServerTime();">
<h1>Ajax Demonstration</h1>
<h2>Getting the server time without refreshing the page</h2>
<form>
<input type="button" value="Get Server Time" onclick="getServerTime();" />
</form>
<div id="showtime" class="displaybox"></div>
</body>
</html>

The main functionality is handled by the getServerTime() function in setting up and

sending the XMLHttpRequest object, and the ajaxResponse() function to
display the time.
view the output page

Whats this business with the random numbers?

Web browsers use caches to store copies of the web page. Depending upon how
they are set up, a browser could use data from its cache instead of making a
request to the web server.

The whole point of Ajax is to make server requests and not to read data from the
cache. To avoid this potential problem, we can add a parameter with a random
string to the URL so that the browser wont be reading data from its cache to satisfy
the request (as then it looks like a different request than previous ones).

This is only necessary if the request method is GET, as POST requests dont use
the cache. (This also seems to be more of an issue with Microsoft Internet Explorer
than with other browsers.)

Sending text back the server

The response stored in XMLHttpRequest.responseText from the server can be

any text that JavaScript is capable of processing as a string.

Thus, you can send back a simple text string as the first example did, or you could
send a string with HTML tags embedded in it. You can process the string using
JavaScript functions (to split it into substrings, add/delete parts of it, etc.).

You could even send back a string that has JavaScript code it in and execute it
using the JavaScript eval() method.

Recall, however, that the responseText property is a read-only variable, so if

youre going to alter it you must first copy it to another variable.
Example with HTML tag

(Change the PHP script to insert HTML tags.)

Example using a table

(As above, change the PHP script.)

The other PHP scripts for the time examples

Heres the script with a simple HTML tag in it.
<?php
echo '<span style="color: red;">' . date('H:i:s') . "</span>";
?>

The output with a table.

<?php
$tr = '<tr style="border: 2px solid;">';
$td = '<td style="border: 2px solid">';
$table = '<table style="border: 2px solid; margin: auto;">';
$table .= $tr . $td . date('j M Y') . '</td></tr>';
$table .= $tr . $td . date('H:i:s') . '</td></tr>';
$table .= '</table>';
echo $table;
?>

XML: a (very) brief intro

XML, the eXtensible Markup Language, is used in many ways, the
most relevant to us being the transfer of structured information.
XML and HTML look similar in many ways and this is because both
are based on SGML, the Standard Generalized Markup Language
established by the International Organization for Standards (ISO).
Like HTML, XML uses tags to denote information but is not limited to
the types of tags that occur in HTML. Tags can be essentially
anything a user likes and are used to define the type of data present
in the document.

XML: a (very) brief intro (cont.)

Heres an example:
<library>
<book>
<title>Programming PHP</title>
<author>Rasmus Lerdorf</author>
<author>Kevin Tatroe</author>
<author>Peter MacIntyre</author>
<chapter number="1">Introduction to PHP</chapter>
<chapter number="2">Language Basics</chapter>
. . .
<pages>521</pages>
</book>
. . .
</library>

See the other notes for some more details/examples.

view other notes

Accessing an XML document in JavaScript

To use an XML document in JavaScript, you must create an object to hold it. This
can be done in the following fashion:
Non-Microsoft browsers:
<script>
var myXMLDoc = document.implementation.createDocument("","",null);
myXMLDoc.load("mydoc.xml");
// other code here
</script>

Internet Explorer:
<script>
var myXMLDoc = new ActiveXObject("Microsoft.XMLDOM");
myXMLDoc.async="false";
myXMLDoc.load("mydoc.xml");
// other code here
</script>

Once weve created the object holding the XML document, we can then use
JavaScript methods to examine it, extract data from it, etc.

The time example using XML

The first change is to make a new PHP script that returns an XML document to the
browser.
<?php
header('Content-Type: text/xml');
echo "<?xml version=\"1.0\" ?>\n";
echo "<clock><timenow>" . date('H:i:s') .
?>

"</timenow></clock>";

After that change (and inserting the new script name into the HTML code), we need
to alter the ajaxResponse function to parse the XML document. That new
JavaScript function is given on the next page.

Note that we need not explicitly create an object to hold the XML document, but
that responseXML (as a property of XMLHttpRequest) is already such an object.

The new Ajax response function

function ajaxResponse() //This gets called when the readyState changes.
{
if (ajaxRequest.readyState != 4) // check to see if we're done
{ return; }
else {
if (ajaxRequest.status == 200) // check to see if successful
{
var timeValue =
ajaxRequest.responseXML.getElementsByTagName("timenow")[0];
document.getElementById("showtime").innerHTML =
timeValue.childNodes[0].nodeValue; }
else {
alert("Request failed: " + ajaxRequest.statusText);
}
}
}

This new response function uses a JavaScript method to access the XML DOM and
retrieve the time string before inserting it into the output display box.
view the output page

A second example (live search)

Well build a live search function. When you typically use a form, you must submit
the data to the server and wait for it to return the results. Here we want to consider
a case where you get (partial) results as you enter data into the input field, and that
these results are updated (almost) instantly.

Note: This example has been adapted from the book JavaScript in 24 Hours by
Michael Moncur.

We use PHP again for the backend. First consider the case where the possible
results are a list of names, stored in a PHP array. As you type data into the input
field, its matched against this list of names, and the (partial) matches are returned
and displayed on screen.

Later, we will see the same type of application, but using PHP to search through
the names stored in a database.

The HTML layout (no JavaScript yet)

<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN"
"http://www.w3.org/TR/xhtml11/DTD/xhtml1-strict.dtd">
<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
<head>
<title>Ajax Demonstration</title>
<style>
</style>
<script>
//
The JavaScript front end will be in here.
</script>
<body>
<h1>Ajax Demonstration of Live Search</h1>
<p>
Search for: <input type="text" id="searchstring" />
</p>
<div id="results">
<ul id="list">
<li>Results will be displayed here.</li>
</ul>
</div>
<script type="text/javascript">

// This sets up the event handler to start the

//
search function.
// var obj=document.getElementById("searchstring");
// obj.onkeydown = startSearch;
</script>
</body>
</html>

view the output page

The PHP backend

<?php
header("Content-Type: text/xml");
$people = array( "Abraham Lincoln", "Martin Luther King", "Jimi Hendrix",
"John Wayne", "John Carpenter", "Elizabeth Shue", "Benny Hill",
"Lou Costello", "Bud Abbott", "Albert Einstein", "Rich Hall",
"Anthony Soprano", "Michelle Rodriguez", "Carmen Miranda",
"Sandra Bullock", "Moe Howard", "Ulysses S. Grant", "Stephen Fry",
"Kurt Vonnegut", "Yosemite Sam", "Ed Norton", "George Armstrong Custer",
"Alice Kramden", "Evangeline Lilly", "Harlan Ellison");
if(!$query) $query = $_GET['query'];
echo "<?xml version=\"1.0\"?>\n";
echo "<names>\n";
while (list($k,$v) = each ($people))
{
if (stristr ($v, $query))
echo "<name>$v</name>\n";
}
echo '</names>';
?>

This PHP script takes the query that will be passed to it, then searches for (case
insensitive) matches to the names in the array.
It passes an XML document back to the calling function consisting of the names
that it finds.

The JavaScript functions

We obviously need the function for creating a new XMLHttpRequest object,

which we will store in a global variable called ajaxRequest.

The search will be handled by setting up a Timeout event, based on entering text in
the input field (using the onkeydown attribute).
var t;

//

public variable for the timeout

function startSearch()
{
if (t) window.clearTimeout(t);
t = window.setTimeout("liveSearch()",200);
}

The liveSearch function is the main calling routine, where we set up the
XMLHttpRequest object, and make the call to the PHP script.

The JavaScript functions (cont.)

Recall that were making ajaxRequest a global variable in the script, so that as in
the other example we can access its properties in the callback function.

function liveSearch()
{
ajaxRequest = getXMLHttpRequest();
if (!ajaxRequest) alert("Request error!");
var myURL = "search.php";
var query = document.getElementById("searchstring").value;
myURL = myURL + "?query=" + query;
ajaxRequest.onreadystatechange = ajaxResponse;
ajaxRequest.open("GET", myURL);
ajaxRequest.send(null);
}
function ajaxResponse() //This gets called when the readyState changes.
{
if (ajaxRequest.readyState != 4) // check to see if we're done
{ return; }
else {
if (ajaxRequest.status == 200) // check to see if successful
{
displaySearchResults();
}
else {
alert("Request failed: " + ajaxRequest.statusText);
}
}
}

The JavaScript functions (cont.)

function displaySearchResults()
// This function will display the search results, and is the
// callback function for the Ajax request.
{ var i, n, li, t;
var ul = document.getElementById("list");
var div = document.getElementById("results");
div.removeChild(ul); // delete the old search results
ul = document.createElement("UL"); // create a new list container
ul.id="list";
// get the results from the search request object
var names=ajaxRequest.responseXML.getElementsByTagName("name");
for (i = 0; i < names.length; i++)
{
li = document.createElement("LI"); // create a new list element
n = names[i].firstChild.nodeValue;
t = document.createTextNode(n);
li.appendChild(t);
ul.appendChild(li);
}
if (names.length == 0) { // if no results are found, say so
li = document.createElement("LI");
li.appendChild(document.createTextNode("No results."));
ul.appendChild(li);
}
div.appendChild(ul); // display the new list
}

The finished product

We add all of the functions (and the two global variables) to the header script
section, uncomment the two lines in the other script section and were good to go!

The fact that the names are in a PHP script allows us to easily add more or delete
some of them. If desired, you could have the search array in a separate PHP file
and include it in the search routine script, allowing you to reuse the same code with
many different lists.
view the output page

Using a database for the live search

Instead of storing the names in an array, we could alter the PHP script to search
through a mySQL database for matches.

The JavaScript need not be changed (except for the name of the script to call).

As before, the PHP script will return the names as an XML document, using
methods for a case-insensitive search on the query string.

A new PHP script is shown on the next page.

<?php
header("Content-Type: text/xml");
echo "<?xml version=\"1.0\"?>\n";
echo "<names>\n";
if(!$query) $query = strtoupper($_GET['query']);
if($query != "")
{
include_once('db_access.php');
$connection = mysql_connect($db_host, $db_username, $db_password);
if(!$connection)
{
exit("Could not connect to the database: <br/>" . htmlspecialchars(mysql_error));
}
mysql_select_db($db_database);
$select = "SELECT ";
$column = "name ";
$from = "FROM ";
$tables = "people ";
$where = "WHERE ";
$condition = "upper(name) LIKE '%$query%'";
$SQL_query = htmlentities($select . $column . $from . $tables . $where . $condition);
$result = mysql_query($SQL_query);
while ($row = mysql_fetch_row($result))
{
echo "<name>" . $row[0] . "</name>\n";
}
mysql_close($connection);
}
echo "</names>";
?>

view the output page

Some cautions

As with any JavaScript element, you cant (or shouldnt) rely upon a users browser
being able to execute JavaScript (some people turn it off on their browsers). (Of
course, there are webpages that ignore this caution.)

Debug carefully and on many different browsers. Ajax uses features that might not
be present in all browsers or they may not operate in the same fashion.

If you can, indicate to the user that something is happening or that something has
changed on the page, otherwise they may not notice it.

Ajax can possibly introduce strange behavior, like the Back button on the browser
doesnt act like it did before (as with any dynamic website), or that if you use some
hidden elements in your page (generated by Ajax), then they will likely not show
up in a form that search engines will recognize.

For the sake of us all (who will be looking at your pages), dont let flashy behavior
be a substitute for actual content, or a well designed layout of your web pages.