Contrary to all that's written about client/server integration, distributed systems, systems integration, and office suites, exchanging data between applications is still mostly by importing and exporting simple text files.
The problem with such data exchanges is that the data in these "flat files" have no meaning. Either a label or the order these data appear in the file will signify that "800-950-8020" is a part number, not a telephone number, and that "99.99" is a price, not a level of purity.
So, while a Web browser can display a catalog page in HTML (Hypertext Markup Language) content, it won't "understand" what in that page is part number, price, telephone number, or whatever. A word processor can also display that page, but it also can't extract the part number and price.
XML changes that. The eXtensible Markup Language (XML), created in early 1998 by the Worldwide Web Consortium, is neither a product nor a platform. Neither is it a revision of HTML. As with so many other information technologies, XML is an enabling technology.
Specifically, XML is a vendor-neutral data exchange language for passing information, not just data.
Like HTML, XML is metadata—data about data. It provides a grammar, a structure, for information. This grammar defines a way to create and embed text-based snippets of code called XML "tags." These tags identify what the data are in an XML document, but not how they are used. That job is up to the software program receiving the XML file. For instance, depending on the application, the string of XML tags and data that constitute part number and price can be displayed as rows and columns in a spreadsheet, used for analysis in a quality control application, or displayed on a Web page for employees and supply chain partners.
Contrast this with HTML. HTML identifies layout; XML identifies data. HTML tells a Web browser how to display the elements in a Web page; XML tells applications what the data elements in an HTML document mean. HTML provides portable publishing capabilities; XML provides a portable database. HTML makes documents human-readable; XML makes documents machine-readable.
Realize that XML alone does nothing; it is merely a data file format. XML requires a browser object, called the XML parser, to interpret the XML metadata (tag), thereby making the individual data elements in the XML file mean something to the receiving application. (As with Web browsers, XML parsers from different vendors have different features—and capabilities).
The Hurwitz Group, an information technology (IT) consulting firm in Framingham, MA, says XML's top three benefits are currently high-end custom publishing, intra-enterprise application-to-application integration, and business-to-business information exchange. The complex automotive repair manuals from OEMs are an example of high-end publishing. While much in these manuals are the same from year to year, each car model year introduces some unique descriptions, making customizations necessary and complex.
Intra-enterprise integration—such as between enterprise resource planning (ERP) and customer relationship management systems—is critical to supply chain and other business collaborations. Because XML enables machine-to-machine communication, it can integrate disparate information systems regardless of platform (mainframe, mini, and personal computers), location (within an enterprise or throughout a supply chain), transmission access method (across a local or wide area network, or across the Internet), or application (including planning, execution, distribution, and customer relationship management). This alone breathes new life into what-some-disparagingly-call "legacy systems."
Most of the current XML software development has been in incorporating XML as an import/export tool that maintains the proprietary data formats in an application. General Motors, for example, is using XML so that Web browsers can display the engineering and quality-assurance data in legacy systems. To do this, Java-based RIO, an XML server from Datachannel Corp. (Bellevue, WA), extracts data from Oracle and IBM databases on Unix servers, among other data sources, formats that data into XML, converts the XML into Web pages, and then posts the pages for Web browsers. Note that both XML and Java are platform-independent; XML formats the data and Java processes it.
Business-to-business information exchange, the third Top-Three XML benefit, is another way of saying e-commerce, electronic data interchange (EDI), and supply chain management technology. Here, XML is the key to electronically exchanging business documents, such as purchase orders and invoices, between trading partners.
That's not saying XML will replace EDI, but it will make EDI far more available, inexpensive, and easier to implement. Case in point: EDI requires matching the data fields in diverse applications. The problem is that a change in one field, such as adding a digit to a part number, often requires remapping everything. In XML, the XML tag (metadata) doesn't care about the particulars of the data element it identifies.
"Virtually any Internet-based communication that has a heavy content and a transactional nature can use XML," says William Swanton, vice president of Manufacturing Strategies for AMR Research (Boston, MA). He expects in the next year or two XML-based semantics and messages created to do certain business processes will become standard-operating-procedure across the general Internet and Automotive Network eXchange (ANX).
But as with conventional EDI, XML has the same caveat that two or more businesses exchanging data must tag the data the same way. Toward that end, many industry groups are tweaking XML to suit their needs, creating industry-specific extensions, namely XML vocabularies and formats. For example, the steel industry is working with Computer Sciences Corp. to develop a set of XML extensions called the Steel Markup Language (SML). SML would include steel industry-specific data formats, such as chemical makeup, weight, and dimensions. The Instrument Society of America SP-95 Enter-prise/Control Integration committee is defining an XML library for passing information between ERP systems, manufacturing execution system, cell controllers, and device controllers.
These industries and more, both in and out of the IT industry, are betting that XML will become ubiquitous. The Web, instant procurement, instant demand fulfillment—all of that is fueling the real-time information, instant visibility, instant communications, and instant compatibility between applications in the supply chain that XML enables.