XML: Cultural problems holding back a killer technology?

It is now almost two years since the world wide web consortium adopted XML as the lingua franca to get to the parts HTML can't reach. While the standard is now adopted by all the major players, it is yet to make a major impact upon the web. Joey Gardiner finds this is due more to people than technology...
Written by Joey Gardiner, Contributor

It is now almost two years since the world wide web consortium adopted XML as the lingua franca to get to the parts HTML can't reach. While the standard is now adopted by all the major players, it is yet to make a major impact upon the web. Joey Gardiner finds this is due more to people than technology...

Last year XML - eXtensible Mark-up Language - entered the technology mainstream. While having received tacit support from the big software houses for a number of years, for the first time XML was at the vanguard - a flag-waving technology, a must have for the big players. Even Microsoft went some way to making its support real through its .NET strategy based on XML protocols. But real adoption continues to be slow. So why are companies not already using XML-based software? The answer is principally because there are still not that many XML solutions out there. The problem turns out to be that XML was never just a replacement for HTML. And it should never have been billed as such, say the evangelists. Rita Knox, VP and research director at Gartner Group, reckons hype shouldn't obscure the real benefits XML will bring. She said: "Has it been overhyped? Have people made outrageous claims? Yes of course they have. It is absurd to think it is going to solve all your problems of data interchange overnight - but it does have enormous potential." The fact is that XML - because of its capacity for holding extensive meta text - is capable of far more than HTML, but as such is many orders of magnitude more complex. XML provides a system whereby information on the web cannot only be published, but can be described, categorised and ordered in minute detail. However, XML does not automatically describe content - all the classification and description must be done by the user. For example, the categories needed to describe information in the chemicals industry are very different from those needed to describe information in the financial services industry. The world wide web consortium W3C - the body responsible for XML - has said it will stay away from this standardisation of categories. It says its part is in creating the framework, not dictating the delivery. This leaves us with a broad standard, that is only made useful through tailoring to specific market needs. And herein lies the reason for the delay. The biggest difficulty seems to be getting industries to work together to provide common standards for the good of all. One example of this customisation is the 18 month old Reuters-led NewsML project. The media giant developed the concept of NewsML and presented it to standards body IPTC (International Press and Telecommunications Council) in October 1999, which now controls its development. For Reuters it made perfect sense, says Jo Rabin, CTO at the firm. "At Reuters we have become increasingly aware of a change in how people receive news," he said. "Given information overload, there is no longer value in getting information to people - the value is now in targeting and editing so people get only the information they need. Also content needs to be in a form where it can be re-packaged, re-branded and re-used in different contexts. NewsML allows all of this." NewsML works by categorising content exactly, so it is easy to retrieve specific information. It also categorises by type of content, allowing news providers to distinguish between text and illustrations, background detail and headlines, and so on - ideal for an aggregator like Reuters. However, Rabin admits that getting the industry to adopt this will not happen overnight: "This is a tremendous piece of work - but it needs consolidation for a year around Version 1 before we can move forward." Although the IPTC has 30 members from the international media community, it does not list players such as Bloomberg, News International and Time Warner, so support from this body cannot guarantee success. To make an impact it has to become a de facto standard as well as a committee-based one. And these problems are being mirrored in other ventures. Guy Gurden, VP at investment bank JP Morgan, chairs a working group designed to progress fpML (Financial Products Mark-up Language) for the derivatives trading market. He agrees that the biggest problem is in getting competitors within a sector to work together: "It is an organisational not a technical challenge. What is taking the time is building effective consortia." The message is clear: if the industry support isn't there, the standard is worthless. And far from being an entirely open standard, the customisation needed to make XML useful means there is overlapping effort, and - a worse case - conflicting XML standards. A classic example of this is the conflict between Microsoft's BizTalk Server and the ebXML project - both aimed at standardising the formats for transmitting messages over the internet. EbXML is being proposed by standards body Oasis (Organisation for the Advancement of Structured Information Systems), whereas Microsoft wants to use its industry weight to make BizTalk the de facto standard. Bob Sutor, a board member at Oasis and a director at IBM agrees that creating industry consensus is the real difficulty. "Technology is not the issue. Co-operation on a global scale with a real mix of people is the real problem," he said. The trouble is that companies who are used to fighting tooth and nail for market share now have to work together to create standards in everyone's interests. And while standards bodies struggle to gain credibility, attempts from vendors - like Ariba, Commerce One and Microsoft - to enforce their own standards look self-serving in the new, open Internet Age. As Gartner's Knox notes, the cultural problems are much bigger than the technical ones, adding: "The biggest hurdle is making people realise we need to share information." It is only with this ethos of sharing information that we can begin to hasten the slothful take-up of XML technologies. While companies are still using standards to gain competitive advantage, progress will inevitably be stalled. However, by next year we should really see whether industries are showing the maturity to enable XML to flourish.
Editorial standards