Presentation-Oriented Mashup Techniques
When you work with a mashup in the presentation domain, you are constrained to an environment that has evolved in a hodge-podge way from processing simple text and graphics pages to one that is on the verge of offering as many or more features as a complex desktop application framework. This messy evolution has suffered because of the mismatch between the free-natured needs of the web environment and the restricted nature of the few primary browser vendors. Nevertheless, standards and best practices are emerging that offer sanity to the mess. The following sections discuss some of the popular techniques and technologies being used to produce mashups in the presentation domain.
Mashing Presentation Artifacts
The easiest form of presentation-oriented mashup involves aggregation of UI artifacts within a web page in a portal-like manner—that is, completely segregated from each other using discrete areas within a single HTML page. In this model, UI artifacts such as gadgets, widgets, HTML snippets, JavaScript includes, and on-demand JavaScript are embedded within an HTML document using layout elements and techniques such as HTML tables and CSS positioning.
When mashing together UI artifacts in a web page using browser layout techniques, each artifact typically resides in its own separate area, as illustrated in Figure 1.6
Figure 1.6 Mashed presentation artifacts
As illustrated in Figure 1.6, a mashup using aggregated UI artifacts references one or more web sites from a web page and retrieves UI code that builds each artifact as a separate component in a separate area on the browser page.
Mashing Presentation Data
A browser page can also build and modify UI artifacts using data retrieved from multiple sources using such data formats as XML, RSS, Atom, and JSON. In this model, the data is retrieved from one or more sites and parsed by the browser, and UI artifacts are created or updated using scripting techniques such as DOM manipulation to alter the resulting HTML document.
Figure 1.7 illustrates the flow of data from service hosts and external sites to a mashup created by aggregation of data in the presentation domain (browser page).
Figure 1.7 Mashed presentation data
This model is more complex than mashing together UI artifacts. In this model, the scripting code that processes the data must be robust enough to handle multiple data formats or restrict the page to accessing services that only support the formats supported by the page. However, since the scripting code will be parsing the data to a fine-grained level, the UI can be created and updated to create a more sophisticated user experience. This model offers more flexibility at the cost of additional complexity.
Using AJAX and the XMLHttpRequest Object
Asynchronous JavaScript and XML (AJAX) is a set of technologies and techniques used for handling data feeds using JavaScript and for creating dynamic web pages with a sophisticated look and feel. One feature of AJAX is the use of JavaScript and CSS techniques to update a UI artifact on a web page without refreshing the entire page. AJAX also features a JavaScript-based HTTP request/response framework that can be used within a web page.
The HTTP request/response framework provided by AJAX is enabled by a component known as the XMLHttpRequest object. This object is supported by most browsers and offers the ability to pass data to an external host and receive data from an external host using the HTTP protocol. Requests can be made synchronously or asynchronously. Asynchronous AJAX requests are made in the background without affecting the web page from which the request is made.
The AJAX framework can send and receive virtually any data format. However, XML-based data and JSON data are the most frequent payloads used for various reasons discussed elsewhere in this chapter. Once data is received it is parsed and applied to the page, typically by manipulating the DOM.
Figure 1.8 illustrates the process through which data flows within a web page using the AJAX framework. As illustrated, when using AJAX, data is received by the XMLHttpRequest object. UI artifacts can then be created and modified using the data.
Figure 1.8 Presentation data mashup using AJAX
Document Object Model (DOM)
Every JavaScript-enabled web page is represented internally to a browser as an instance of the W3C Document Object Model (DOM). DOM is a platform-independent and language-neutral object model for representing XML-based documents that allows programs and scripts to dynamically access and update the content, structure, and style of a document.
The HTML DOM is part of the official DOM specification that defines a standard model for HTML documents. The HTML DOM facilitates accessing and manipulating HTML elements in a given web page. The HTML DOM presents web page as a node-based tree structure containing elements, attributes, and text. Every element on a web page (for example, div, table, image, or paragraph) is accessible as a DOM node. JavaScript allows the manipulation of any DOM element on a page dynamically. This allows you to perform such operations as hiding elements, adding or removing elements, and altering their attributes (color, size, position, and so on).
Listing 1.1 presents a simple HTML document.
Listing 1.1. Simple HTML Document
<html> <head> <title>A simple HTML doc</title> </head> <body> <p> This is a simple HTML document. </p> <img id="image1" src="https://example.com/image1.png" width="250" height="350"/> </body> </html>
Listing 1.2 is an example of a JavaScript function that changes the width and height of the image (<img>) element with an id of image1 in the preceding example:
Listing 1.2. JavaScript Manipulation of DOM
function changeImageSize() { var anIMGElement = document.getElementById("image1"); anIMGElement.width = "400"; anIMGElement.height = "300"; }
As shown in Listing 1.2, the DOM can be accessed by the global "document" variable. With a reference to the document variable, you can traverse the nodes of the DOM to find any element by name or id.
Extensible Markup Language (XML)
XML (eXtensible Markup Language) is a specification and standard for creating self-describing markup languages. It is extensible in that it allows you to define your own elements. It is used extensively in data transformation and integration frameworks to facilitate the transfer and integration of structured data across disparate systems and applications. XML is used in many web-enabled environments as a document annotation standard and as a data serialization format.
Listing 1.3 is an example of a simple XML document.
Listing 1.3. Simple XML Document
<?xml version="1.0" encoding="UTF-8"?> <contacts> <contact> <name>John Doe</name> <address1>123 anywhere st.</address1> <address2>Apt 456</address2> <city>Yourtown</city> <state>CA</state> <zip>12345</zip> <country>USA</country> </contact> <contact> <name>Jane Doe</name> <address1>456 S 789 W</address1> <address2>Suite 987</address2> <city>Mytown</city> <state>NY</state> <zip>54321</zip> <country>USA</country> </contact> </contacts>
Presentation-oriented mashups consume XML and XML derivatives returned from service hosts and web sites. Once XML is received, it is parsed and applied to a given web page. As XML is parsed, DOM manipulation techniques are usually applied to update UI artifacts.
JavaScript Object Notation (JSON)
JSON (JavaScript Object Notation) is a simple, string-based, data-interchange format derived from JavaScript object literals. JSON is very easy for users to read and write and for JavaScript engines to parse. Strings, numbers, Booleans, arrays, and objects can all be represented using string literals in a JSON object.
Listing 1.4 is an example of a simple JSON object:
Listing 1.4. JavaScript Object Notation (JSON) Object
{ 'contacts': [{ 'name':'John Doe', 'address1':'123 anywhere st.', 'address2':'Apt 456', 'city':'Yourtown', 'state':'CA', 'zip':'12345', 'country':'USA' }, { 'name':'Jane Doe', 'address1':'456 S 789 W', 'address2':'Suite 987', 'city':'Mytown', 'state':'NY', 'zip':'54321', 'country':'USA' }] }
Presentation-oriented mashups also consume JSON objects returned from service hosts and web sites. Once a JSON object is received, it must be parsed in order to apply to a given web page. Since JSON is pure JavaScript, it is easily parsed using standard JavaScript utilities. As with XML, DOM manipulation techniques are usually applied to update UI artifacts once a JSON object is parsed.
Sidestepping the Browser Security Sandbox
Perhaps the biggest challenge in doing a presentation-oriented mashup is contending with the browser security sandbox, which is in place to keep sensitive information secure. To protect against malicious scripts, most browsers only allow JavaScript to communicate with the host/server from which the page was loaded. If a mashup requires access to a service from an external host, there is no easy way to access it using the XMLHttpRequest object.
When an attempt is made to access a host/server external to the host/server from which a web page was loaded, an error similar to the following will be encountered:
Error: uncaught exception: Permission denied to call method XMLHttpRequest.open.
Therefore, a mechanism is needed through which services can be accessed without violating the browser security sandbox. JSONP provides one solution.
JSON with padding (JSONP) or remote JSON is an extension of JSON where the name of a JavaScript callback function is specified as an input parameter of a service call. JSONP allows for retrieving data from external hosts/servers. The technique used by JSONP is referred to as dynamic or on-demand scripting. Using this technique, you can communicate with any domain and in this way avoid the constraints of the browser security sandbox.
Listing 1.5 is an example of using on-demand JavaScript to retrieve data from an external site.
Listing 1.5. On-Demand JavaScript
function retrieveExternalData() { var script = document.createElement("script"); script.src = 'http://www.example.com/aservice?output=json& callback=aJSFunction'; script.type = 'text/javascript'; document.body.appendChild(script); }
Listing 1.5 illustrates how to dynamically add a <script> tag into a page by manipulating the DOM so that the page can load and call another web site. The <script> tag executes on-demand and makes the service request to the site specified. If the service response is in JSONP format the JavaScript interpreter converts the response object into a JavaScript object. When a script element is inserted into the DOM the JavaScript interpreter automatically evaluates the script. JSONP responses wrap JSON data in a JavaScript function call using the name of the callback parameter. The JavaScript function is then called and any objects defined in the JSONP response are passed.
The following is an example of an evaluated JSONP response:
aJSFunction({ "item 1":"value 1", "item 2":"value 2" });
As shown in the preceding line of code, the response returned from the service is formatted as a JSON object wrapped in a JavaScript function call with the callback parameter name. The script is evaluated and the JavaScript function is called, completing the service request/response interaction.