Columns

XML's demi-decade

• By Uche Ogbuji
• March 28, 2003

XML turned five years old in February. Surely, no other technology has so swiftly worked its way into every technical conversation. Windows went through more than a decade of desultory releases before Windows 95 vaulted it into the mainstream. Linux took a little less time from Linus's first, famous message to its current inevitability. Java is perhaps the closest rival in this measure of rapid adoption, but even it had less of a meteoric rise.

"Few of the technologies advanced by the XML community are in themselves revolutionary," I said about XML in 1998 in "LinuxWorld." "The power of XML lies in the fact that so many vendors and users are finally agreeing on practical standards for such models."

This agreement and cooperation has continued to a remarkable degree even to this point. I have always believed in the power of diversity to further science, and I think XML has enjoyed the great benefit of a variety of participants.

"Programmers pounced on XML to replace their frequent and expensive invention of one-off data formats, and the specialized parsers that handled these formats," I wrote in the "Intel Developer Journal" in 2001. "Web architects seized [on] XML as a suitable way to define content so that presentation changes could be made cleanly and easily. Database management experts adopted XML as a way to exchange data sets between systems as part of integration tasks. Supply-chain and business interchange professionals embraced XML as an inexpensive and flexible way to encode electronic transactions. When so many disparate disciplines find themselves around the same table, something special is bound to happen."

In this column I have argued that XML injects fresh and very productive ideas into the discipline of software development, offering a fundamentally different view of data to programmers. I have controversially held that the "XML text way" (which I've called the "bohemian" way) leads to less-risky development and maintenance. I have also discussed how in at least one case --the incorporation of strong data typing into XML -- a conflict has opened up between bohemians and "gentry," those who support the greater import of concepts from traditional programming. Recently, I and other bohemians have founded the XPath NG project to try to improve the most important XML processing spec while maintaining a focus on text.

Despite the overall accord in XML, bohemian/gentry is just one battle that will shape the next five years of XML. A similar debate is heating up the Web services community: systems with strong bindings to programming languages vs. "document/literal" systems that stick to more textual fundamentals.

Another battleground is intellectual property (IP). Standards groups have been polarized and founded over IP stances. OASIS mandated royalty-free (RF) standards from the start; the WS-I was created by those who wished to enforce IP in Web services; and the W3C ended up mostly in the RF camp. Other XML-related technologies have seen narrower battles; for example, the dust-up over Sun's IP related to XPointer, and the claims of a patent prospector over RDF and RSS. A new outcry has been raised by a batch of patents, including material by Microsoft, that seem to lay claims over rather fundamental aspects of XML technology.

The more subtle, but perhaps most important, debate is over how to evolve XML. Everyone knows of some shortcoming in XML, and there have always been alternative formats to XML, but these have mostly been on the sidelines and recognized as being distinct from XML itself. Recently there have been more proposals to directly supersede XML while adopting all its terminology. Notably, there is the call of SOAP proponents for an XML "profile" that bans DTDs and that maybe even alters syntax to allow easier nesting of XML documents. I and others believe that such tinkering with the bedrock of XML -- syntactical interoperability -- is dangerous and counterproductive. Nevertheless, this conflict may determine the basic shape of XML in the next five years.

To read more columns by Uche Ogbuji, click here.