What SGML Includes That HTML Does Not

SGML probably has many capabilities from which HTML could benefit. At the same time, some proponents of HTML want to keep many of SGML’s features out because they fear that HTML will become too complicated and cluttered. Table 17.1 describes the versions of HTML.

What HTML 3.0 Adds

Version 3.0 of the HTML DTD adds the following features. Some are not terribly significant, but others are important.

•  Easier hot zone shaping

---

Tip:  
A hot zone is a vectored portion of an image that is mapped with a specific hyperlink. Your mouse cursor changes shape while over the hot zone to indicate to you that it is linked to another resource. When you click on the hot zone, your browser jumps to the URL associated with that hot zone.

---

---

• See “DSSSL,” p. 419

• See “Panorama Pro,” p. 453

---

•  The CLASS attribute for easier use with Stylesheets or DSSSL Lite

•  The ALIGN attribute for mathematical equations

•  The CLEAR attribute for directing text flow

•  Entities for standard icons (%HTMLicons) and math symbols (%HTMLmath)

•  Math and equation structures to facilitate external equation macro languages like TeX

•  Entities for special symbols other than math symbols

•  Powerful table support

•  More extensions, including FILE (for uploading files to a server) and AUDIO (for playing and recording audio samples)

Likewise:

•  The <BASE> element lets you de-reference URIs that get in your way

•  The <STYLE> element lets you override stylesheets and have more flexibility

•  The <RANGE> element lets you address marked ranges more easily

•  The SRC attribute has been added to the <SELECT> element for support with graphical selection menus

•  More flexibility is available for linking with the %linkType, %linkName, and %ToolBar entities; you can have links that are almost two-way

•  More Web pages can take advantage of the %heading entity and the <TITLE> element

•  The new <BANNER> element for banners at the top of the window does not scroll with the rest of the page; this helps when you are working with corporate logos and disclaimers

•  Tighter SGML declarations appear throughout; older and more dangerous structures are marked depreciated

In the pipeline for the next revisions are:

•  Support for client-side scripts that work with programming objects

•  Table entry capabilities, so that visitors to a Web site can enter tables

•  More flexible formatting options for TEXTAREA fields

Table Support. Tables are one of the most difficult challenges in SGML. That is because they demand a balance between format and content. If you markup the information just to satisfy the formatting requirement, you lose track of the meaning of the content. If you markup the information strictly according to content to facilitate database searches and manipulation, then you must redesign the table structures each time the meaning of the content changes. It is very difficult to strike the balance between the formatting and content requirements of tables in SGML.

---

• See “Handling Tables,” p. 421

---

However, the recent CALS table model attempts to strike such a balance by giving the SGML processing system—and external search engines as well—a way of locating information in SGML tables. The key to the solution lies in relating the cells of a table to columns, and the columns to values in table headings. By assigning attributes to table formatting elements that correspond with the meaning of the content, SGML markup structures can help search mechanisms locate information in variable content tables. This powerful type of table model is included in HTML 3.0.

Equation Support. Because HTML was originally developed for science researchers, you would think it should already contain support for equations. But the fact is, equations are difficult. Not only do you require numerous public entity sets with special symbols, but you must also have an automated way of displaying them in the proper relation to each other. Nowadays, there are many public entity sets with math and equation symbols, but displaying them in the proper format still requires either extensive DTD structures, an external equation editor, or both.

---

• See “Handling Math and Equations,” p. 424

---

The proposed HTML 3.0 revision adds still more math symbols and equation entities, but it also accommodates arrays for storing macros that can execute formatting tasks. This way an external language like TeX could be used to format complex equations for Web pages instead of displaying equations as graphic images, or not at all, which is so common today.

Client/Server Document Structures. The File and Audio extensions, as well as the table entry capabilities, suggest more interchange of individual document structures between server and Web client. As the difficulties of this process become resolved, paradigms for server-side SGML document validation can possibly emerge and be tested. This means that when you are using your SGML browser to research documents, servers can validate their local SGML documents and resolve conflicts—like missing DTDs, entity sets, and public declaration files—before your SGML browser returns an error to you.