SGML: Another XML article

Another XML article

From w3c-sgml-wg-request@w3.org Mon Apr 28 11:16:34 1997
Date: Mon, 28 Apr 1997 09:11:11 -0700 (PDT)
From: Todd Freter <Todd.Freter@Eng.Sun.COM>
Subject: Another XML article

 ---------------------------------------------------------------

Folks,

Jon Bosak is traveling on business, updating Sun Microsystems' staff in Grenoble about XML. He faxed me an article from Le Monde Informatique about the latest XML developments (as of April 18, 1997) and asked that I translate it and post it to the group. Enjoy.

As usual, the strengths belong to the author Cyril Dhénin, and the errors are all mine.

-Todd.


[See the original version of the article in French: "CONFÉRENCE ANNUELLE DU WORLD WIDE WEB: Netscape reconnaît XML", by Cyril Dhénin, in Le Monde Informatique n718 - 18 avril 1997, or archive text copy without the (essential) cartoons.


Netscape Acknowledges XML

(Le Monde Informatique, April 18, 1997)

A dramatic switch on the stage of the World Wide Web Consortium: After having pretended unawareness of its existence, Netscape has just acknowledged the attraction of XML -- a second document-description language for the Web at the heart of the W3C, already supported by Microsoft.

HTML no longer has a monopoly on the Web. At the heart of the World Wide Web Consortium (W3C), which brings forth Web protocols, there has just emerged a second language, XML, which stands for eXtensible Markup Language.

Already supported by Microsoft and Sun, this protocol was still pooh-poohed a few days ago by Netscape. When we interviewed him on this topic (see Le Monde Informatique of March 21, 1997), Mike McCue, a director inside the firm for advanced technologies, replied that he didn't see the appeal of a new document-description language.

Those comments contrast markedly with comments made at the annual Web conference by Carl Cargill, Netscape's director of standards. This key player in the firm effectively admitted that Netscape no longer saw any obstacle to adopting XML in its products. This 190-degree turn was justified, so said Mr. Cargill, by the realization that XML does not compete with HTML. And also because, if XML and HTML are both SGML derivatives, their positioning is different (see sidebar, below).

In order to develop software for navigating and developing XML documents, software developers can adopt two tracks: Either they build these tools from the ground up, or they use one of the SGML engines already available on the market. Because the Web's document description languages (HTML and XML) are knockoffs of SGML, using such an SGML engine would save a remarkable amount of time. Two companies that are quite cozy with the W3C are rich in this technology today: the Canadian SoftQuad and the French Grif. In fact, these two companies, in the weeks to come, are probably going to be wooed by Microsoft and Netscape...

For the moment, in the race for SGML smarts, Microsoft appears to have a significant lead. In particular, Microsoft recruited one of Grif's cofounders, Jean Paoli, who is now Data Product Team Manager for Internet Explorer and a co-developer of the XML specifications ... The emergence of this language reveals a tight convergence of interests between Microsoft and the W3C. On one hand, the Seattle [sic] firm wouldn't be disappointed to end up with the HTML monopoly that, historically, remains associated with Netscape. On the other hand, the SGML community and the W3C seem determined to promote a second language for the Web. That's why Microsoft has sped up on XML while Netscape has ignored it -- at least, as long as it could ignore it.

Cyril Dhénin

[sidebar]

Two languages for different needs

In order to understand the respective positioning of HTML and XML, it's important to keep in mind that they are both products of SGML, a generic standard for document description, heavily used in the realm of EDI (electronic data interchange). HTML is an SGML application whose simplicity and light weight fit well with the constraints of the Web. But for SGML users, HTML suffers from a great flaw: It is a finite language which does not allow the introduction of customized instructions related, for example, to an industry. This HTML flaw has led the W3C to sponsor efforts to develop an extensible language that is an abridgment of SGML and not an ad hoc application. In practice, HTML should remain the preferred language for sites where complex documents or documents tightly linked to the disciplines of some professional domain are not published. By contrast, for vertical applications, XML should be the natural choice.