Special Edition Using SGML:SGML’s Emergence on the World Wide Web

SGML Resources as Non-HTML Document Instances

The advent of SGML browsers for the World Wide Web has introduced a wealth of native SGML documents that can now be viewed in their intended format and structure. Although HTML is useful and workable, you can use other SGML DTDs just as easily with an SGML-capable browser.

• See “Collaboration on SGML Standards,” p. 523

SGML-capable browsers open the door to a wealth of document content. You don’t have to settle for HTML’s limited support for tables or equations for your Web documents; you can write your own customized DTDs to support whatever document features you need. The same goes for other special features—graphics, music, Braille translation, newspaper columns, and so on. Simply make sure that your DTDs and documents are valid SGML by parsing them completely and by making them public so that the SGML browser can locate them and process them. If there are any applicable output specifications or stylesheets, they must be made publicly available as well. Also make public any additional processing software (on all possible platforms) for handling special file types, such as music or Braille.

This type of SGML document is new on the Web. SGML documents have been available via FTP for some time, but you had to browse them offline. Only now is it becoming possible to easily browse them online as if they were HTML documents. The next section discusses the variety of SGML documents available.

SGML Resources Available on the Web

After you obtain an SGML browser, take a tour of SGML documents in their native DTDs. Native DTDs mean non-HTML DTDs, and non-HTML document instances on the Web are rather new. This tour starts at:

http://www.ncsa.uiuc.edu/SDG/Software/Mosaic/WebSGML.sgml

Figure 18.3 shows what this home page looks like. It has links to a variety of document instances for SGML DTDs.

Fig. 18.3 The SGML Tour of the Web is a starting point for a tour of different SGML document types—not just HTML.

By following these links, you can see that when you break the mold for HTML, different types of document features become easier and more feasible. Figure 18.4 shows you how well tables can be handled by using a DTD other than HTML.

Fig. 18.4 This SGML Web page illustrates the advantages of not being limited to one SGML document type. The tables here are done according to the AAP DTD.

Figure 18.5 shows a document that HTML could not handle as well as another SGML DTD. The equations would challenge an HTML author.

Fig. 18.5 HTML could not handle this SGML document type well. The equations are from the AAP DTD.

Publishing on the World Wide Web is still a new phenomenon, and much experimentation is going on. Only in the last several months have people been customizing the backgrounds of home pages with different wallpapers. This feature is not strictly SGML, and it is not supported by all browsers. However, it shows that authors want to experiment. As more SGML browsers appear on the market, more authors of HTML Web pages are becoming interested in authoring more than one type of SGML document.

SGML Resources Available on the Web in the Future

The creativity evident on the World Wide Web leads to valuable resources. Judging by some of today’s projects, the SGML resources of the future will be abundant.

• See “Current Collaborative Projects on the Web,” p. 519

Too many projects are in progress to discuss in detail, but some of the most down-to-earth ones have been the most helpful. TEI, for example, has provided useful DTDs that are currently in use. The Modern English Electronic Library at the University of Michigan uses the TEI P3 DTD, and all those wonderful classics are now available as SGML documents.

Scientific and Technical Data

Scientific literature will become more accessible. These documents can be more challenging to make accessible than straight text. DTDs have been made and are continuously being upgraded. TEI is again an example of this. Novell and Intel have made numerous technical documents accessible to the Web; they are both part of the SGML Tour that Panorama starts you off on (refer to fig. 18.3.) These documents require full SGML—even more so than literary documents—because of their intense use of tables, equations, and graphics.

SGML will enable even more scientific and technically oriented graphics than HTML currently does. The trouble with graphics is that there are so few formats that translate easily for all platforms, including PC, Mac, and UNIX. With SGML’s use of stylesheet and resource grabbing from locations around the Web, these compatibility issues might be minimized.

Stylesheets

Another bright prospect for the future of SGML has to do with alternative stylesheets for home pages. There is a large appetite for graphics on Web pages. A Web user today has little choice in how to view Web pages that are full of graphics. He can turn off inline graphics and see no graphics, or he can leave it on and wait while his browser loads enormous quantities of graphics on some pages. Web authors under SGML can configure multiple stylesheets for the same page, each offering the Web client a different viewing alternative. The same page can be configured with heavy graphics, medium graphics, or few graphics, depending on the Web client’s preference and the connection speed. Every Web page then becomes several pages; each page has its own stylesheet. Both the client and the author are happy.

The debate between standardization and innovation will continue to escalate, but SGML’s use of multiple DTDs and stylesheets may help. Because SGML allows more flexibility than HTML, ambitious Web authors can attempt more daring feats that do not defy standards. Netscape, for example, does not adhere to the HTML standard—and it is wildly successful. Stylesheets and flexibilities with DTDs would seem to answer many of the demands for further creativity. The special Netscape-enhanced tags, like <BLINK> and others, could be supported under full SGML without invalidating the HTML standard, and various stylesheets could support many of the other graphic-oriented Netscape innovations, all without violating the international standard. As full SGML becomes more popular, the flexibility that SGML offers can only be exploited by following standards carefully. Otherwise, accessibility is compromised.

Table of Contents