XML makes its mark

By Evan Hansen and Paul Festa
Staff Writer, CNET News.com
February 10, 2003, 4:00 AM PT

What if every bit of data in every computer included instructions about its content that would allow any other computer to interact with it?

Such interoperability could unleash amazing new automation and efficiencies in information systems, and spawn a powerful new service-driven computer industry. For example, software might be written that would allow a carmaker to instantly change its parts orders across all of its suppliers to meet a sudden rise in demand for a specific model.

Although we're not there yet, hundreds of the world's best and brightest technology companies are betting on some version of that vision becoming a reality. Marketed relentlessly under the buzz phrase "Web services," it is a still largely unrealized goal. But there are signs of real progress, if not revolution.

One key benchmark of success has been the steady adoption of a coding language known as Extensible Markup Language (XML), which was approved as a standard by the World Wide Web Consortium (W3C) five years ago today.

XML began as a tool for publishers to describe the contents of documents exchanged over the Web, but it quickly became popular as a technique for describing any kind of data. The language provides terms used to define a Web document's tags--that is, the elements of the document that describe its various pieces--and the relationships between them. Developers at either end of a data exchange then agree to use a common set of tags.

Hundreds of companies now support XML, which is fast becoming the Web's new lingua franca, laying a foundation for a coming wave of Web-based services.

Five years after XML's birth as a W3C recommendation, CNET News.com caught up with some of the people closest to its genesis to gauge the successes, failures and coming challenges for a key technology that has come to underpin the Web.

 XML is the alternating current of the Web and of distributed computing in general. Just as electrification didn't take off in earnest until (the) industry agreed on the alternating standard we use today, so too with XML. In fact, this pattern is a familiar one from history: proprietary technologies, followed by a speculative boom and bust, leading to the establishment of underlying standards that usher in the golden age of the technology. It happened with electrification, the railroads and standard track gauges, and it's happening now with distributed computing.

XML's inherently extensible nature, combined with its applicability for both messages and documents, has made it the foundation for the explosion of standards built upon it, many of which fall under the Web services banner. As a result, we're seeing a sea change in the industry, as virtually all software vendors are moving to support these standards. ZapThink believes that once this transition period concludes, the golden age for distributed computing and the Web will begin. So on this fifth birthday of XML, I'd like to say: The best is yet to come. 

 The five years since XML was released have seen XML become the lingua franca of the Web. But this universal embrace has not always been accompanied by a clear understanding of what XML can and cannot do.

What XML cannot do is to magically solve the problem of data interoperability. XML just provides a framework within which interested groups can work out agreements about the vocabularies and data structures to be used in a given domain. The widespread adoption of XML has created a wonderful infrastructure of standardized tools and products to support the creation and implementation of such agreements, but deep down, the job of semantic definition requires the same grinding committee work that standards groups have been engaged in for more than a century.

On the other hand, there is relatively little awareness of one big thing that XML can do: It can play an essential role in freeing its users from the big-vendor hegemony that has ruled the computer industry for the last 50 years. The ability of user communities to develop their own data formats is a powerful force for freedom from vendor control.

Consider electronic commerce. Bridging the gap between rich and poor economies is a global imperative. Businesses of all sizes must be brought into the EDI framework currently occupied by the Fortune 500. Doing this economically will require royalty-free data standardization and inexpensive software as well as vendor support. A combination of XML-based standards and technologies is now converging to accomplish this goal. The ebXML standards provide a free, coherent, easily implementable infrastructure for trade that maps to existing EDI systems; UBL provides standard business messages; Gnome, Linux and Java provide a free, vendor-neutral computing platform; open-source products such as ebxmlrr and OpenOffice provide free registries and office productivity tools; style sheets and open-source page formatters allow the large-scale output of printed business documents; and commercial products like the Sun ONE Secure Trading Agent are coming online to provide vendor support for trading partner agreements and secure messaging over the free Internet. The convergence of these elements will enable the entry into electronic commerce of most of the world's businesses. I'm proud that Sun Microsystems continues to play a leading role in this movement--a role that it adopted when it organized and led the creation of XML itself.  

 Pinpointing XML's influence is a hard challenge. How has the phone influenced you? XML is everywhere, and indeed that is what is amazing about it. XML is how we create Web content, how we integrate computers, how we define vocabularies and languages for communicating between companies and how we package those messages. It is also in our databases and how we get data into and out of those databases. My non-XML friends always say the most important thing about XML is that it is everywhere.

For example, Contivo was able to help Agilent develop and deploy a common information model for integrating a wide variety of business systems using a standardized XML vocabulary defined by OAGI. This would have been impossible without the XML-based convergence.  

 XML was created to be a foundational technology, and in that goal it has far surpassed the expectations of those of us who were there at the beginning. Everyone knew that we were creating something that would simplify the development of applications and the management of information, but no one knew how far it would really go. Five years later, we still don't know. XML has spread its reach into almost every industry, and it's just continuing to expand out from there. Virtually every type of structured content or data collection worth representing has been, or is currently being, developed as XML. Millions of people use XML every day, and most important, many of them don't even know it. This is only going to increase as XML matures and becomes the enabling technology for Web services and single-source publishing.  

 We knew XML could be a universal solvent for electronic data, but what's been exciting is watching it be used in new ways we hadn't even thought of at the time. Much of the initial adoption of XML has been in support roles rather than being a visible Web publishing phenomenon; application integration messages and B2B transactions are meant to be consumed by machines, not eyeballs. These uses of XML may not seem exciting, but they have increased the pace and quality of commerce in ways we couldn't have expected five years ago. Even directly on the Web, however, technologies such as graphics and multimedia have benefited from XML specifications such as SVG (Scalable Vector Graphics) and SMIL (Synchronized Multimedia Integration Language), since these open specifications have allowed for greater choice and more competition.

Years ago, I heard Dan Connolly use the "Pig Farmer's Markup Language" as his hypothetical example of an XML tag set. Soon after, he discovered that such a thing was already in the works for real--leading me to believe are no merely hypothetical XML languages anymore.

Some people are concerned that XML just forces us to contend with hundreds or thousands of splinter languages, and doesn't simplify anything at all. But to me, these many efforts to define XML vocabularies are proof that it works. As long as communities are empowered to define their own tags, and as long as each effort is open to all stakeholders, the ability to communicate will only increase.  

 I remember the day we started XML; it's hard to believe we are now celebrating its fifth anniversary. XML took the planet by surprise, and the news today is that there is no news when two different software platforms use open-standard XML to exchange data seamlessly over the Internet.

I would say that the most exciting premises of XML in my eyes are the possibilities of changing the way users interact with computers through a document metaphor. You may ask yourself why it is XML that is changing the world and not another technology? I believe that the reason is the core design of XML that provides, with its semistructured model, a unified vision of information between data and documents. There is fundamentally no reason why documents created by humans cannot be processed by databases and then searched and reused in other ways or systems. The extent of people's imagination is phenomenal and, for the first time, XML empowers them to express their ideas into documents that are semantically rich.

Vendors are building back-end systems and messaging layers with Web services that support XML. It is hard for me to overstate how firmly I'm convinced that XML documents generated by millions of desktops will become the diamonds of the Internet--preciously exchanged, reshaped, examined in their multiple facets, shared and aggregated by the numerous online services both on the Internet and inside and across the enterprise.  

 XML is the foundation for the evolving architecture of the Web. When XML 1.0 was released by the W3C five years ago no one knew how far-reaching its effects could be. Two years after XML's release, VoiceXML was born to bridge the gap between the Web and the phone. VoiceXML stands as a classic XML success story. It allows businesses to bring the power, flexibility and quality of their Web applications to the phone. Today, Fortune 500 companies depend on VoiceXML to power thousands of phone systems and answer millions of calls every week, creating unprecedented customer satisfaction and saving companies millions of dollars.

In just three years, VoiceXML has achieved widespread industry adoption, making it the most broadly supported and implemented voice standard in the world. The open-standard framework of XML has made all of this possible, and in the specific case of VoiceXML, will continue to drive innovation on the phone in the years to come.  

 In the past five years, XML has gone from an obscure format for technical manuals to a central part not just of the World Wide Web, but of modern computing and business.

Applications of XML range from configuration files through to remote procedure calls, from desktop menu definitions to chat protocols. With every new usage of XML, the value of all the interoperable XML tools, both proprietary and open source, is increased.

Of particular interest on the World Wide Web is the success of XSLT, the XML way of transforming documents. This XML-based language has proved useful as glue, as middleware, connecting among other things databases, text documents, Web servers, Web browsers and styles. It's also finding a place as part of Web services, helping to process business transactions.

Two early design decisions with XML were to minimize the number of optional features, and to require strict error-checking. The result of these decisions has been that there is a very high degree of interoperability between tools. Another decision was to use Unicode as the document character set: Work on accessibility and internationalization are helping to make the World Wide Web truly a place for all people.

Five years ago, if you had suggested to a programmer that a configuration file be stored in SGML, you probably would have gotten either a blank stare or a hostile reaction.

In stark contrast, XML is used today by the Gnome desktop on Linux, by the Jabber chat protocol, and is even central to the upcoming release of Microsoft Office. We've made it. We're mainstream.  

 Yuri Rubinsky, one of the founders of SoftQuad, had reserved a slot for a W3C working group entitled SGML On The Web. The plan was to make SGML documents available over the World Wide Web. Yuri, sadly, died in January 1995. But the slot at W3C remained, and (Sun's) Jon Bosak got permission from Tim Berners-Lee to activate the working group. Jon picked nine (later 10) people to be on the working group. They were chosen because of their backgrounds. Most had extensive, real-life SGML experience spanning many years. Many also had HTML experience. Tom Magliery, for example, had worked on the original NCSA Web browser, Mosaic, and I had built HotMetal. That small group wrote the first version of the spec in just a few months. Somewhere in those months the name, XML, had been created and the SGML ERB (Editorial Review Board) had been renamed as the XML (working group).

We each brought different skills and experiences to the working group. If you ask almost anyone on the working group, they could tell you a story where they had already invented XML before the working group even met. I have such a story. Steve De Rose even wrote part of a book about something similar to XML. Collectively we put these ideas together and created XML. If I had a quality that distinguished me from the others in the group I think it was that I always tried to represent the average, maybe even naive, user. I tried to apply the lessons learned from HTML. The biggest lesson was that you can do a lot with something that is really simple. SGML wasn't simple. XML is.

We were confident that XML would be useful. We had dreams that it would be successful. But the reality has been much greater than I think any of us had imagined.

XML has certainly changed the computing world. More than just Web applications, I think the biggest payback has to be in terms of content reuse. Governments and private industry have changed the way they create content and in turn, the way they create documents and Web sites. As more organizations realize these benefits, we are going to continue seeing significant changes in the way people think of publishing.  

 Happy birthday XML! The past five years have really borne fruit for XML and some of its most important applications, especially Web services. The dramatic uptake of standardized ways of representing information has had a significant impact on the way companies think about the information they produce and the applications they share. Part of the evidence of this influence is that, as analysts who now completely focus on XML and Web services and the way they are changing enterprise architectures, we have spoken with more than 500 software vendors and end users--none of which now thinks that producing data in proprietary formats or with proprietary interfaces is tenable in the long term. The pervasiveness and widespread adoption of XML has in fact changed the economics of technology adoption--while in the past it might have been economically feasible to work in a vendor-proprietary, closed environment, the reverse is now the case. It is less risky and more cost-effective to adopt open standards for connecting companies and systems. As this trend continues, XML promises to be as ubiquitous as TCP/IP.

XML and Web services have also been the singular bright spot in what has proven to be a significant IT spending downturn. Why is that? Primarily because we see XML and Web services technologies as helping companies save costs, make the most of their existing technologies and leverage existing skill sets. The result? XML and Web services spending really started to take off in 2002, and promises (more in) 2003. Case in point? Our analyst firm, ZapThink, which has focused entirely on XML and Web services, grew more than 800 percent in revenue this past year. This is the same year that saw Hurwitz close its doors and other analyst firms suffer. Perhaps our own firm is evidence of the growth and adoption of XML and its importance to enterprises.  

 I came to XML via a path not often reported: through academia. Because my background is in Germanic philology (Old Norse, Old and Middle English, Old and Middle High German language and literature), my initial interest in computers was in finding better ways for computers to help the scholarly study of textual material. Before you can study texts with computers, though, you have to represent texts in electronic form.

SGML--the parent of XML--was the best available language for encoding texts in a structured fashion--but software was both hard to find and hard to build. So what I most wanted in XML was something that would provide the same rich semantics as SGML, but would be easier to write software for.

The most dramatic consequence of XML, for me, has been the flood of XML-aware software: XML parsers, XML-aware programming-language libraries, XSLT processors, XSL FO processors, XML awareness in databases--and not least, XML-aware Web browsers. Most of them are intended by their developers to be used for commercial applications, but as it happens, the most difficult problems faced by scholars working with difficult texts invariably correspond to similar problems faced by commercial software users, and vice versa. And where commercial applications don't directly address a problem, the ubiquity of XML programming tools means it's easier to fill the gaps.

It is now much, much easier to set up systems to deal intelligently with richly marked-up text, which means great things for organizations both commercial and academic, and for the preservation of our cultural heritage.  

 Netscape is strongly committed to supporting XML and related Web standards because they form the foundation for building the 21st century interactive Web. We are working to ensure that the XML foundation remains open, secure and innovative. Netscape recognized the importance of XML early on and built XML into the heart of the Netscape 6 and 7 browser series. Browsers can now dynamically connect to Web services for specific data transactions without a fresh page load, resulting in a fast and dynamic Web experience.

Netscape and Mozilla's XML-based XUL (Extensible User Interface Language) is also an example of how XML can be extended to provide cross-platform user interface development where developers can customize the browser or create entirely new applications. In addition to using XML for core capabilities such as its user interface, we pioneered advanced uses of XML, such as MathML, an XML-based markup language for representing mathematical data. Web services based on XML protocols such as SOAP (Simple Object Access Protocol) are also an important paradigm shift for browser technology, enabling the Web to become as dynamic a medium as any desktop application. Over the coming years we will see more and more applications that will take advantage of XML and Web service capabilities within the browser in order to become faster, easier to use, more informative and more maintainable.