Three Little Letters but One Big Leap?

By David Streitfeld
Washington Post Staff Writer
Tuesday , September 19, 2000 ; G25

The XML programming language may be taking the Internet to a new level, but its many fans haven't yet settled on a mundane way to describe to the uninitiated exactly what it does.

It's the glue that knits the Net together, says one. A bar code for information, says another. It's the lingua franca of electronic commerce and content, says a third. It's an Esperanto for computers, says a fourth, referring to an attempt to merge the main European languages into a universal form of communication.

If anything, such labels understate the fervor with which XML is being touted. Networking guru Craig Burton has said XML will do for this era what calculus did for the Renaissance: Both are dynamic new approaches that render the previous system outmoded, if not irrelevant.

David Turner, a product manager at Microsoft Corp. specializing in XML, goes even further.

"The introduction of XML is in many ways like the creation of writing in the evolution of language," he said. "People had spoken language for a long period before they got to the point of inventing writing. But as soon as they did, they were able to make huge steps forward."

XML, which stands for extensible markup language, is technically a metalanguage, a language that describes other languages. It is a loose cousin of hypertext markup language, whose development laid the groundwork for the Web. But while HTML is a fine tool for writing Web pages, it's static. HTML presents a fixed snapshot of data. XML merely structures it, allowing for much more fluidity.

With XML, Web sites can exchange data much more easily, a process that greatly facilitates business-to-business commerce. The research firm Gartner Group has estimated that by next year, 70 percent of business-to-business transactions will use XML.

Let Microsoft Chairman Bill Gates, whose new bet-the-company .Net strategy is founded on XML, take a crack at it: "Say you want to take a look at where in the country you want to sell a particular kind of product, or look at certain kinds of outlets for a product. That's obviously something where today you'd have to go to the Yellow Pages, you'd have to go to different mapping sites, you'd have to go look up the catalogues of the different stores, and all you want to do is see a map of where a certain type of merchandise can be bought."

XML will change that, Gates told his audience at a recent industry conference: "XML messages would be sent out to the various sites, and a very rich view would be synthesized for you that brings that information together."

Microsoft, just like every other high-tech company, is gambling on a future where all devices will be compatible: Refrigerators will tell the supermarket you're almost out of milk, music can be zapped from PC to rec-room stereo to car to cell phone, electronic address books will call you up at work to remind you of that dental appointment. XML is the key to all that, too.

The developers of XML, a working group from the standards-setting World Wide Web Consortium, started with ambitious goals, but ones that didn't have much to do with business-to-business applications.

"XML was invented by a bunch of people with a track record in the publishing industry," said Tim Bray, a member of the development team. "We thought we were building the document format of the future. The enemy, explicitly, was Microsoft Word, Frame, Quark--all these fragile proprietary binary file formats that lock up the inventory of human knowledge. Nobody was thinking about B2B."

Bray noted that he had been "publicly and amusingly sneered at" by a Silicon Valley executive he declined to name, who scoffed: "You guys look like a bunch of geniuses, and all you did was say, 'Make the tags balance.' "

HTML, in other words, had already introduced the notion that the Net should be based on a stream of text with markers called "tags" embedded in it. What XML did was insist that the tags be used in an orderly and complete way.

"He has a point; there's really not that much technology in XML, and even less that's new," Bray conceded. "But there are a few touches that, while small, have been crucial."

They include:

* XML is text, not binary. When computers ship data, they can do it either as characters people can read, or as a bunch of binary 1s and 0s only another computer can read. "Computers often disagree about binary representation," Bray said. "They always agree on how to represent the letter 'A.' What XML proves is that doing it as text always wins."

* Superior error-handling. HTML doesn't require exactitude, which is good: It empowered the nontechnical individual to become an Internet publisher. But in B2B applications, you don't want the receiving application guessing what the sender meant. It should be clear what a purchase order says, otherwise you might end up with 14,000 bananas instead of 14. "XML has no wiggle room. If the rules aren't followed, the receiving application is required to emit an error message and stop," Bray said.

* Internationalization. "The programmers of the world are discovering that, by gosh, the Internet isn't entirely in North America, and yes, with XML you can include in your purchase transaction sourced from Tokyo your name in kanji, or in Cyrillic from Moscow, or whatever," Bray said. "The days of the Anglophone Internet are numbered, so this is a good thing."

If XML was originally targeted at certain companies' proprietary products, that hasn't stopped virtually every company from embracing it. At an XML conference earlier this month, presentations were made by Microsoft, International Business Machines Corp. and Software AG, Europe's largest system software vendor.

"There was a high degree of portability between the presentations," said Turner, who gave the Microsoft talk. "They all had the same message about how this was the next generation of solutions based on open standards protocols."

Such agreement is highly unusual in the computer world, which has been plagued by competing standards over things as simple as character sets--the ways computers present text. Forced to choose between developing open standards that anyone could use and proprietary standards that everyone would with luck be forced to use, high-tech firms generally have opted for the latter.

Just because there are open standards, of course, doesn't mean there isn't a fierce battle for customers, as tech firms from Microsoft and IBM on down try to help other businesses convert. In Seattle, a venture capital firm was set up in January solely to fund XML companies, echoing what the high-profile Kleiner Perkins group did a few years ago for start-ups using Java technology.

"XML is phase two of the Internet," XML Fund founder David Pool said. "If XML didn't exist, the Internet would just be a bunch of magazines. XML makes it a platform for computing. By turning the Web into a giant database, XML will have a bigger impact than the browser did."

One of Pool's dozen investments so far, in XYZFind Corp., caught the eye of Zona Research. "Who cares about some little techno-geek company being funded by some relatively small VC fund?" Zona wrote in a report earlier this month, and then answered its own question:

"In this case, probably entities like Metropolitan Life, Nasdaq and AT&T on one hand, and all of the Web-based comparison-shopping engines like CNET, ZDNet, MySimon, Street Prices and CompareNet, among others."

The reason for Zona's excitement is found in XYZFind's new-generation searching mechanism, which can scan XML-encoded data. Basically, XML makes searches smarter, allowing them to distinguish between the same number used in different ways--$10,000, for example, and 10,000 miles.

For the e-commerce user, Zona said, the net result will be "more precise, intelligent returned data for his commercial queries."

Impressive as all this sounds, it's worth remembering that the Internet has always been fertile ground for hype. And while no one thinks XML is equivalent to, say, "push" technology--a briefly hot system that quickly chilled--some say the enthusiasm is getting a trifle extreme.

A key XML developer, David Megginson, rejected the calculus and written language comparisons, settling instead for an analogy with the standardization of the rail gauge in the late 19th century, which allowed one railroad's boxcars to be passed on to another's.

"XML isn't the discovery of anything," he said. "It's not like the invention of the Web itself, which allowed people to do things in a whole new way. It's just a compromise everyone can live with for structuring data. It's an argument--how were we going to represent low-level data?--that had to be resolved before we could go further."

© 2000 The Washington Post Company