XML and Information Systems XML

XML and Information Systems XML can help to enable effective communication in information systems by providing differing programs with a common format for data interchange. Many programs use binary data dumps of their internal data structures as file formats. These data dumps are not easily portable between programs and are therefore poorly suited for use in systems where many differing programs need to communicate.

XML is not special or unique as a data format, but it is a standard with broad industry support, which makes it ideal for use in a system that needs to communicate with arbitrary or unknown programs. One practical application of XML as a communications format is the open-standard Jabber instant messaging protocol. Another is the increasingly common practice of using Javascript to update web pages without reloading them, retreiving XML content asynchronously over HTTP.

XML, as a standard enables developers of such systems to ignore the low-level details of sharing data and focus on their applications. One of the easiest ways for a programmer to save data to a file or stream it across a network is to simply dump the data structures, unaltered from memory. Such data dumps are easy for the program that created them to read, and quite difficult for anything else. This has traditionally been a popular technique among Windows developers; Unix developers have traditionally preferred text-based file formats.

(Raymond 2003) Traditional Unix text formats, while easy to parse, even without knowing the internal details of the program that created them are not formally standardized. Generally, a program must by specifically designed to read and write another program's file or stream format. XML provides a generic tool for doing so, often supported by an external library. The XML syntax should look familiar to any Lisp programmer; Lisp property lists have exactly the same structure as XML documents, and have served the same purpose since long before XML was intented.

(Graham 2002) If Lisp were the dominant language for software development, XML would probably not exist, however, Lisp is primarily associated with artificial intelligence programming and computer science education; it is only used for end-user application development by a few highly successful companies.

Writing a library for any language to export data as a Lisp property list would be just as easy as doing the same for XML, however, XML is the de-facto standard, so it is the most sensible thing to use when communication with other applications is a potential issue. Jabber is an open-standard for instant messaging. It has many user-visible features in common with commercial services such as ICQ and AOL Instant Messenger, but lacks the central server that those systems use.

Instead, Jabber is organized more like email, with each organization running its own server. As with email, multiple client and server programs exist which impliment the Jabber protocol. (Jabber Technical Overview 2005) The use of an XML-based data interchange format simplifies the development of compatable software. In an older Unix application, a specialized text-based format would have been used to send messages over the network. While such a system would work perfectly well, using XML means developers are not responsible for the error-prone task of writing a parser.

If the current email system had been developed more recently, it too would probably use XML. The various add-on features email mesages have developed, as well as the common practice of adding extra headers to email are an ideal fit for the extensibility of XML. An XML-based system would generally be less error-prone to impliment, but does add bulk compared to the simplicity of the original email message format, a disadvantage for use on small devices such as cellular phones.

For Jabber, using XML means the developers only had to design data structures, not a format for transmitting them. It also means that any program that needs to understand part or all of the Jabber protocol only needs to understand XML. The benifit to developers is obvious to anyone who has worked on third-party clients for popular instant messaging services. The protocols for such services are often difficult to distinguish from line noise.

Even if such services were not trying to make their protocols difficult to reverse-engineer, Jabber's would likely still be easier to work with because of its XML basis. Software which supports XML can easily handle supporting new XML document tyes, such as the Jabber protocol, and a raw dump of the traffic would look familiar to anyone who has built a web page: tag-based markup intended to be human-readable. When the World Wide Web was new, many predicted that it would replace traditional desktop applications.

Java applets were usually the means by which this was to happen, though in reality they have much in common with regular desktop applications. Using normal web pages as a user interface to an applications has significant disadvantages relative to a normal desktop application, most of which are related to the fact that the user must load a new page every time any significant interaction with the application occurs. Such page loads are inconvenient for users, and have discouraged the use of the web to replace traditional desktop applications.

With the introduction of the Javascript XMLHttpRequest object, web application design has become more like the desktop. XMLHttpRequest allows a web page to communicate with a server using XML. Javascript already had the ability to update content on an already-rendered page. The addition of the ability to communicate with a server means that web applications can truely start to behave more like desktop applications. The technique has been given a name: Asynchronous Javascript And XML, or Ajax, which has been widely adopted.

(Garrett 2005) Writing an Ajax application involves significantly more complexity than a traditional web application, but has very large potential usability benifits for the user. For an Ajax application to work well, several elements are required: standards-compliant HTML and CSS, a client-side Ajax engine written in Javascript, and server-side processing to respond to the requests. XML is used as the format for bi-directional communication between the web page and the server. Updates are made by the Javascript Ajax engine without having to reload the page, making for a much.