InformationWeek

It turned out that Extensible Markup Language, or XML, was at the heart of the controversy.

We'll get to the details of XML in a moment, but first, let's clear up the "proprietary" confusion: XML is not a Microsoft format. It was originally proposed in 1996 by the World Wide Web Consortium, or W3C; (see http://www.w3.org/XML/) and has been moving forward ever since. When finished, XML will be a full, open, worldwide standard.

The "proprietary" charges stem from the fact that, ironically, Microsoft is far ahead of Netscape Communications and Opera Software in adopting XML. IE4 was the first browser to ship with an XML parsing engine built in; IE5 ships with an engine that conforms pretty well to the current W3C specs. Netscape users will have to wait at least until for Mozilla 5 for a similar level of built-in conformity (see http://www.mozilla.org/rdf/doc/xml.html). Opera is silent on the subject of XML. To the casual eye, it may appear that Microsoft is doing something proprietary simply because it's the only one shipping a highly XML- compliant browser. (See http://msdn.microsoft.com/xml/c-frame.htm#/xml/default.asp) But in this case, and to Microsoft's credit, it's just that the company is far ahead of its competitors.

OK, what's the big deal with XML? Why should we care? Well, for one thing, XML is far, far more flexible than HTML. The latter is mostly about layout, but XML isn't about layout per se; it's about the data--the content.

XML lets you almost completely separate the content of a Web or intranet page or document from its layout scheme. For example, you can have a page or document set up to display data that's stored in a separate file. That separate data file can be updated at will (entirely or granularly) without touching the actual Web page or document itself. No Java applets or special server-side includes or other tricks are needed at all; XML handles everything, simply.

Or, XML data can be embedded within a page or document in something called "data islands." The data in the island can be sorted, searched, or otherwise manipulated from inside the browser (or other XML-enabled application) without having to make additional data calls to a server. On an intranet, offloading these search/sort/, etc., functions to the client reduces the load on the server and the LAN; on the Web, it reduces overall traffic and improves apparent responsiveness by avoiding network latency issues. (The storage and transmission of XML data can be enhanced further through compression; XML data is plain ASCII with many repeating elements, so it squeezes down very nicely.)

XML allows virtually any data element to be defined in almost any way. At the most trivial level, this means Web authors can define tags on the fly, and those tags will be understood and acted upon correctly and automatically by XML-enabled applications everywhere.

But more significantly, it means that all XML data can be self-describing to almost any desired level of granularity. One benefit is that searches can be faster and more precise through the use of narrowly focused, high-precision data tagging. Instead of having to wade through an entire document with a brute-force search, XML searching will let you zero in on exactly what you're looking for--more wheat, less chaff. XML also enables searching across multiple, incompatible data sources, and it can combine the results easily and (to the end user) seamlessly.

Curiously, XML has two faces. When you play by XML's rules, you gain enormous freedom in tagging, how you handle data, where data is located, and so on. But you really must play by the rules: A conforming XML application isn't supposed to guess at what half-baked XML code means or take a stab at presenting badly formatted data. It's supposed to simply refuse to display questionable material. This ensures that intranet and Web XML applications will do what they're supposed to do, and nothing else; accuracy and repeatability go way up.

In a way, XML is as much of a leap over basic HTML as today's HTML 4.0 is over, say, HTML 1.0. But XML and HTML are not necessarily in competition: The W3C even has a spec for what it calls XHTML, the Extended Hypertext Markup Language, which blends HTML 4.0 and XML 1.0 (see http://www.w3.org/TR/xhtml1/).

All this barely scratches the surface of XML; the links in the text above will let you dig as deeply as you wish to learn more.

But make no mistake: XML is not proprietary. It is enormously powerful and flexible, and it will be at the heart of tomorrow's enterprise applications.