XML creator castigates binary plan

A technology standards-making body proposes a way to make XML files smaller and, in turn, easier for computers to process, but Jean Paoli, one of the mark-up language’s creators, isn’t fond of the idea.

The World Wide Web Consortium (W3C), which mulls interoperable technologies for the Web, recently recommended researching a new method of turning XML files into binary documents. The group said binary XML would make it easier for television set-top personal video recorders (PVRs) to source programming information from broadcasters, and help companies exploring for oil and gas deposits collect data on fuel wells and seismic conditions.

But Paoli, Microsoft Corp.’s senior director of XML architecture and an XML founding father, figures the W3C is going about it all wrong. While he recognizes a need for XML optimization, he doubts the W3C’s suggestion for a single, over-arching binary XML schema makes sense.

“I believe there are some places where you need binary, but I believe it’s going to be multiple binaries, depending on the scenario,” Paoli said while visiting Toronto last month. “If you want to have two databases talking to each other quickly, I don’t know what kind of binary solution would fit, but I’m telling you it’s going to be different from another scenario where you need XML optimization for cell phones.”

Paoli said optimization standards already exist. It’s easy to zip XML files for instance. As well, the W3C recently approved XML-Binary Optimized Packaging (XOP), which uses Multi-purpose Internet Mail Extensions (MIME) to efficiently link XML files with images.

Thanks to these pre-existing standards, the binary XML plan doesn’t make sense, Paoli said. “First, use what’s available today. Second, optimize per scenario. And never forget Moore’s Law. What seems today to be a problem, just wait a couple of years.”

During his Toronto visit, Paoli described some of the challenges he and other XML pioneers faced while defining the language — a project that started with the GML movement in the early ’80s, progressed through SGML and culminated with XML becoming a veritable W3C standard in 1998.

Paoli, who started working on XML as a simplified version of SGML in the ’90s, said one of the biggest hurdles had to do with convincing influential backers to get on board.

“The challenge wasn’t really creating XML. When we created it, no one cared about it but Microsoft, in terms of the big companies. Even in Microsoft, I had to explain why they should bet on me. I had to demonstrate that it was useful.” Paoli said he tackled SGML all wrong at the beginning. “I personally made a mistake for 10 years before the XML story by focusing too much on the user interface….I would go to a bank in France, for instance, and say, ‘Why don’t you use SGML?’ I’d convince them with a proposal. Now what would they do with it? Go find a database that knows how to deal with it. Well, there were none. There were some people in the SGML community trying to create the back-end systems, but you needed the big guys — the IBMs, the Microsofts.”

Paoli, whom Microsoft recruited in late ’95, turned the development around with XML, putting the user interface at the end of the development list. “For the first five years of XML we focused on the back end. It was on purpose….We put it in BizTalk, Messaging, SQL Server 2000.” “In ’99, I knew it was time to move to my original passion — user interfaces,” Paoli explained. “That’s how we engineered it: back end first, the messaging and then UI.” XML has made its way into many Microsoft products, include front-end ones like the Office suite.

Paoli said the standard’s proliferation provides opportunities for developers, and more than a handful of people are taking advantage. A recent survey showed that a million developers are building apps atop Office programs, and 300,000 of them are specifically building XML apps, Paoli said. “I have a prediction,” he said. “After seeing all these numbers, I believe in 2010, 75 per cent of documents worldwide will be created in XML. I’ve been thinking about it for a long time. If we do a really good job it might go to 80 per cent.”

Paoli said he’s happy to see his brainchild, XML v.1.0, become a global standard, but he’s not keen on Microsoft using the follow-up iteration, v.1.1, just yet. It’s a matter of interoperability — once you have it, you don’t mess with it. “I don’t want to break this for XML 1.1, or some binary we don’t know anything about.”

QuickLink 052562

Related Download
Virtualization: For Victory Over IT Complexity Sponsor: HPE
Virtualization: For Victory Over IT Complexity
Download this white paper to learn how to effectively deploy virtualization and create your own high-performance infrastructures
Register Now