神马文学网城市胡同: 技术
来源：百度文库 编辑：神马文学网 时间：2024/04/28 03:48:43
Adding Ajax to a Website
By kevinwu on January 3, 2009 2:11 PM | No Comments | No TrackBacksCreating a dynamic, user-friendly website interface is simple and straightforward
by Andrew Turner
Introduction
Modern websites and web-applications appear drastically different from sites on the web 5 and 10 years ago. Tools like GoogleMail, BaseCamp, and TiddlyWiki have revolutionized the general concept of what a webpage can do and how users interact with it. The days of clicking on a simple hyperlink to be taken to a new page, or sitting and waiting for a form submission are rapidly dwindling.
The technology driving these sites is not really new, but their application and use has only recently become widespread and supported by a majority of web browsers. Furthermore, many web developers feel daunted by the rapid pace of the changing techniques and don't have a clear understanding of how the technologies are implemented and used.
One of the most revolutionizing of these technologies has been dubbed AJAX (or Ajax depending upon whom you ask). Ajax is responsible for dynamic page content, marking database entries, and in-line text-editing without the need for page-reloads or large, complex plug-ins like Flash or Java.
The goal of this article is to teach you the basics of Ajax and demonstrate that it is not as difficult a concept as it may first appear. In reality, Ajax is simple and easy for any web developer to add to their new or already existing site.
What is Ajax?
AJAX is an acronym for: Asynchronous Javascript and XML. The most important concept of AJAX is the "asynchronous" part. Asynchronous communication means that commands do not need to wait for a response. By contrast, synchronous communication requires the command to wait for a response before continuing. An example of synchronous communication is a typical hyperlink; the user clicks a link, and then waits while the resulting page is requested, returned, and displayed. An asynchronous example may be having a contact name and phone number lookup with dynamic autocomplete with names already in the database. For application developers, it may be useful to think of synchronous communication as modal, while asynchronous is non-modal.
Javascript is the client-side scripting language that has been used to implement the input, output and server-response handling. Because the code is client-side it is fast and scales up with increasing usage. The last part of the AJAX acronym is XML (extensible Markup Language), which is used in the response to encapsulate information. By using a structure like XML, the client can parse the tree for specific data without having a predetermined order of the data. As we will discuss, XML or Javascript are not required to implement "Ajax" in a site.
Ajax uses several other technologies and functionality to work. XHTML and the Document Object Model (DOM) allow Javascript to dynamically modify a webpage. CSS (Cascading Style Sheets) are not necessary, but are typically employed to provide for easy layout and design of a webpage and allow the Ajax functionality to work on the data, and not the view.
The reason the technology is referred to as either AJAX or Ajax is because of the blurring between the concept, and the implementation. Ajax (non-acronym) has become the terminology associated with the ability to dynamically modify a webpage or backend content without requiring a page reload, while AJAX (acronym) is the specific implementation of Ajax employing Javascript and XML. The term Ajax was coined by Jesse Garnett of AdaptivePath (see resources) as a better name than the previously used "Asynchronous JavaScript + CSS + DOM + XMLHttpRequest". The technologies were all originally combined by Microsoft for developing their Outlook Personal Information Manager (PIM) web application interface.
Why use Ajax?
While the web has inarguably drastically changed the way a computer user works, to date they haven't been able to fully replace, or even work entirely in tandem with, desktop applications. To clarify, a desktop application is software that must be installed on a user's computer and is run in a self-contained window/context. By contrast, a web application operates primarily within a user's browser and is not required to be installed on a machine. This provides users access to the application and associated data from any computer using a suitable browser.
However, with the advent and widespread use of technologies such as Ajax, users can now complement, or even replace, their desktop applications with a web application. Many users are now switching to reading their email in GoogleMail, storing their documents and notes in TiddlyWiki, reading their RSS news via Gregarious, or working with colleagues in BaseCamp.
Adding Ajax to your own web site or web application provides a much smoother, and rich user experience. Furthermore, Ajax websites more closely imitate their desktop counterparts, allowing users to interact and understand the user interfaces in a similar way.
Ajax is also a relatively straightforward and simple technology to provide in a website. Developers may quickly become confused by all of the terms, techniques, and options. However, at its core, Ajax is quick to setup and begin using, and completely flexible for whatever the developer and site requirements need. Ajax can be used for features such as inline form validation, database queries, content editing, drag-and-drop, page updating, and many others
Setting up the framework
Parts of Asynchronous Communications
In order to understand the essential parts of an Ajax framework, we will discuss the necessary parts of asynchronous communications. The parts are split up by Client, the user's browser, and Server, the website hosting server.
Client: Create a request object
BClient: Assign a response handler
Client: Send a query to the server
Server: Receive the query, and perform operations
Server: Send the response to the client
Client: Handle the server response
Figure 1: Asynchronous communications allow a user to continually interact with the browser, and provides dynamic updating of the web site
The client requests to the server can happen continually, updating the web page or application on each response received. Each request happens separately from the interface, allowing the user to continue to view and interact with the web page.
However, it is a good idea to only support a single asynchronous command at a time as the response may affect the interface data. If multiple asynchronous requests are supported, you must be careful to handle potential conflicts due to user interaction with outdated data.
Create a request object
The first thing to do is to create a constructor that will build a client-side request object. A request object is responsible for wrapping up the actual request, response handler, and state of the request.
Remember that this Javascript code is being run on the user's desktop browser. Therefore creating the request object is the one place where browser specific code is required. In this case, Microsoft's Internet Explorer uses an ActiveX object as the request object, whereas the other browsers all support an XMLHttpRequest() constructor call. We can interrogate the browser to find out what type it is and create the appropriate object. This function is universal for any Ajax use.
AjaxFramework.js
// request object constructor
function createRequestObject() {
var ro;
var browser = navigator.appName;
if(browser == "Microsoft Internet Explorer"){
ro = new ActiveXObject("Microsoft.XMLHTTP");
}else{
ro = new XMLHttpRequest();
}
return ro;
}
You should now create a global request object that will be used by the client for all future communication.
AjaxFramework.js
// global request object
var http = createRequestObject();
Assign a response handler and handle the response
Our second step is to assign a response handler. A handler is the function that will be called when the request comes back from the server to the client's computer. This function is responsible for verifying the state of the answer, and parsing the response as appropriate. This function is implemented on a project specific basis. It needs to know what the expected response from the server looks like and how to place that response back into the user's browser document.
AjaxFramework.js
// callback function that handles any state changes of our request to the server
function handleResponse() {
if(http.readyState == 1){
// request loading
document.getElementById("status").innerHTML
= "requesting...";
}
else if(http.readyState == 4) {
// request complete
if(http.status == 200) {
// OK returned
var response = http.responseText;
// Add more advanced parsing here if desired
document.getElementById("responseArea").innerHTML
= response;
}
else
{
document.getElementById("status").innerHTML
= "error: " + http.statusText;
}
}
}
The first thing the handleResponse function does is check the current state of the request object. If the object is loading (1), then the user is alerted to this, or if the request is complete (4) then we handle the response. This example just puts the response text (use responseXML for an XML response from a server) into our document's responseArea.
Send a query to the server
Now that we have setup the request object structure, as well as the state handling function, the next step is to create a function that our webpage will be calling for each outgoing request. This function could either accept information via an input parameter, or retrieve user input by querying the document.
Once the user input is received, we create a GET request to a URL. It is important to note that due to security concerns the request can only be made to a server that is hosting the webpage. The domain name must be exactly the same as the request URL, if there is a preceding www. to the domain name. As usual, however, there are some fairly straightforward work arounds for getting external data for your Ajax requests. Several options will be discussed.
This example demonstrates using a REST input (parameters passed via the URL), but other remote query and command options are also possible. Furthermore, the open command supports passing a username and password to the server for accessing protected services.
AjaxFramework.js
// function for filling out and sending a request - called by the actual webpage
function sendRequest() {
var query = document.getElementById("queryInput").value;
var queryURL = "http://localhost.com/service.php?q=" + query;
http.open('GET', queryURL);
http.onreadystatechange = handleResponse;
http.send(null);
return true;
}
We have now completed the necessary parts of our Javascript code to handle creating, sending, and receiving an asynchronous request through a client's browser.
Server handling of the request
The client makes a request to some service or page that is served on the same domain as the original webpage. This service for this example is expecting a value passed via the URL in the GET parameters. The response can be well formed XML, or simple text that will be parsed by the client's browser as discussed above in the handleResponse() function.
service.php
$query = $_GET['q'];
$response = some_service_handling($query);
echo $response;
?>
This server page just passes the query onto another php function and then echoes the response. Since our Ajax request from the browser has made a GET request, this operates like any normal opening a page in a browser. However, instead of the page showing up in a window, it is handled by the client's handleReponse() function.
Using a remote service
As we mentioned earlier, security does not allow the Ajax, specifically XMLHttpRequest, to call another domain in the GET URL. The way around this is provide a locally served wrapper to the remote service. We can parse and pass on each of the incoming parameters. Also, many hosting services don't allow a URL to be opened via the fopen() command, so this example uses curl to make a request to a server. The subsequent response is read by the local server and then returned to the calling Ajax function.
remote_service.php
$remote_params = "";
foreach($_GET as $key=>$value)
{
$$key = $value;
if($value != '')
$remote_params .= "&".$key."=".$value;
}
$remote_url = "http://remotehost.com/remoteservice.php?";
function get_content($url)
{
$ch = curl_init();
curl_setopt ($ch, CURLOPT_URL, $url);
curl_setopt ($ch, CURLOPT_HEADER, 0);
ob_start();
curl_exec ($ch);
curl_close ($ch);
$string = ob_get_contents();
ob_end_clean();
return $string;
}
$content = get_content ($remote_url.$remote_params);
echo $content;
?>
This example makes no checks on the incoming request. The query and parameters are passed directly onto the remote service. In a real application, it would be responsible to do some basic parameter checking before passing on the request to someone else's hosting service.
That said, it is still a means by which to provide asynchronous services in your own website. You should also be aware that making these remote calls may have longer response times. While this situation is an excellent reason why an asynchronous interface provides a better user experience, users may be left wondering if their request was just lost. Therefore, you should, when appropriate, let the user know that the request is pending in some way.
Furthermore, it would be possible to setup a timeout timer for each request that would call abort() on the request object if the request took too long.
Supporting non-Javascript functionality
This framework will generally work for any modern, Javascript capable browser. However, not all users are using Javascript capable browsers, and other users may have disabled Javascript. Therefore, it is advised that your site support a non-Javascript version of your interface. At the very least, alert the user that they will not be able to use all of the functionality of your web application or page.
To provide a non-Javascript interface only when necessary, your page should use the