XML was developed in the mid 1990s under the leadership of Sun Microsystems employee Jon Bosak. Jon was looking for ways to use the Internet for more than just information delivery and presentation. He wanted to create a framework that would allow information to be self-describing. That way applications could guarantee that they could access just about any data. That in turn would clear the path to intelligent data-sharing between different organizations. And that in turn would allow more and much better applications to be written and increase the demand for servers to run them on. Well, that last part isn't a goal, but it's certainly a great side effect for anyone in the computer hardware industry.

Information access might not sound like a problem in these days of web publishing, but it used to be a significant barrier. The web is still not a good medium for arbitrary binary data or data that is not text, pictures, or audio. A few years ago, every hardware manufacturer had a different implementation of floating-point hardware, and the formats were incompatible between different computers. If you had a tape of floating-point data from an application run on a DEC minicomputer, you had to go through unreasonable effort to process it on another manufacturer's mainframe. IBM promoted its EBCDIC (Extended Binary Coded Decimal Interchange Code) convention over the ASCII (American Standard Code for Information Interchange) codeset standardized in the rest of the Western world. People who wanted to see their printouts in Japanese resorted to a variety of non-standard approaches.

By storing everything in character strings, XML avoids problems of incompatible byte order (big-endian/little-endian) that continue to plague people sharing data in binary formats. By stipulating Unicode or UTF encoding for the strings, XML opens up access to all the locales in the world, just as Java does.

XML makes your data independent of any vendor or implementation or application software. In the 1960s, IBM launched a transaction processing environment called CICS. CICS was an acronym for “Customer Information Control System.” When a site used CICS, after a while it usually became completely dependent on it, and had to buy large and continuing amounts of hardware and support from IBM in order to keep functioning. People used to joke that it was the customer that was being controlled, not the information. But it was no joke if you were in that position. Modern software applications cause the same kind of single-vendor lock-in today. XML goes a long way to freeing your data from this hidden burden. But note this key point: just because something is published in XML does not make it openly available. The DTD and semantic meaning of the tags must also be published before anyone can make sense of non-trivial documents.

So XML makes it possible for otherwise incompatible computer systems to share data in a way that all can read and write. XML markup can also be read by people because it is just ordinary text. So what new things can be done with XML? XML opens up the prospect of data comparisons and data sharing at every level on the web. If you want to buy a digital camera online today, you might spend a few hours visiting several retailer websites and jotting down your comparison shopping notes. With XML, you take a copy of the merchants' product datasheets and run an automated comparison sorted in order of the product characteristics that matter most to you. Even more important, if you're a business that needs to buy 1,000 digital cameras for resale, XML lets you put this business-to-business transaction up for bid in an automated way.

Two things have to happen for business-to-business (B2B) automated XML bids and comparisons to occur.

Neither of these is outlandish. Various industry groups have already started to cooperate on common data descriptions. The best known are RosettaNet for electronics, and Acord for insurance. The development community is also working on XML-based protocols to let software components and applications communicate using over HTTP. One contender here is SOAP—Simple Object Access Protocol—from IBM, HP, Microsoft, and others.

XML gives content providers a data format that does not tie them to particular script languages, authoring tools, and delivery engines. XML supports a standardized, vendor-independent, level playing field upon which different authoring and delivery tools may freely compete.