Introduction to HTML 3.0

HyperText Markup Language (HTML) is a simple markup system used to create hypertext documents that are portable from one platform to another. HTML documents are SGML documents with generic semantics that are appropriate for representing information from a wide range of applications. HTML markup can represent hypertext news, mail, documentation, and hypermedia; menus of options; database query results; simple structured documents with in-lined graphics; and hypertext views of existing bodies of information.

HTML has been in use by the World-Wide Web (WWW) global information initiative since 1990. The HTML 3.0 specification provides a number of new features, and is broadly backwards compatible with HTML 2.0. It is defined as an application of International Standard ISO ISO8879:1986 Standard Generalized Markup Language (SGML). This specificiation will be proposed as the Internet Media Type (RFC 1590) and MIME Content Type (RFC 1521) called "text/html; version=3.0".

How to participate in refining HTML 3.0

The process of refining HTML 3.0 into a formal standard will be carried out by the IETF HTML working group. The World Wide Web Organization is continuing to develop a freeware testbed browser for HTML 3.0 ("Arena") to encourage people to try out the proposed features. The discussion list for HTML 3.0 is www-html with html-wg reserved for use by the IETF working group for detailed matters relating to the formal specification. The process for developing HTML 3.0 is open, and anyone who is interested and able to contribute to this effort is welcome to join in.

Note: make mailing list names into hypertext links to their archives and add info on how to join these lists

HTML 3.0 Overview

HTML 3.0 builds upon HTML 2.0 and provides full backwards compatibility. Tables have been one of the most requested features, with text flow around figures and math as runners up. Traditional SGML table models, e.g. the CALS table model, are really complex. The HTML 3.0 proposal for tables uses a lightweight style of markup suitable for rendering on a very wide range of output devices, including braille and speech synthesizers.

HTML 3.0 introduces a new element: FIG for inline figures. This provides for client-side handling of hotzones while cleanly catering for non-graphical browsers. Text can be flowed around figures and you can control when to break the flow to begin a new element.

Including support for equations and formulae in HTML 3.0 adds relatively little complexity to a browser. The proposed format is strongly influenced by TeX. Like tables, the format uses a lightweight style of markup - simple enough to type in by hand, although it will in most cases be easier to use a filter from a word processing format or a direct HTML 3.0 wysiwyg editor. The level of support is compatible with most word processing software, and avoids the drawbacks from having to convert math to inline images.

The Web has acted as a huge exercise in user testing, and we have been able to glean lots of information from the ways people abuse HTML in trying to get a particular effect; as well as from explicit demand for new features. HTML 3.0, as a result, includes support for customised lists; fine positioning control with entities like &emspace; horizontal tabs and horizontal alignment of headers and paragraph text.

Additional features include a static banner area for corporate logos, disclaimers and customized navigation/search controls. The LINK element can be used to provide standard toolbar/menu items for navigation, such as previous and next buttons. The NOTE element is used for admonishments such as notes, cautions or warnings, and also used for footnotes.

Forms have been extended to support graphical selection menus with client-side handling of events similar to FIG. Other new form field types include range controls, scribble on image, file upload and audio input fields. Client-side scripting of forms is envisaged with the script attribute of the FORM element. Forms and tables make for a powerful combination offering rich opportunities for laying out custom interfaces to remote information systems.

To counter the temptation to add yet more presentation features, HTML 3.0 is designed (but doesn't require) to be used together with style sheets which give rich control over document rendering, and can take into account the user's preferences, the window size and other resource limitations, such as which fonts are actually available. This work will eventually lead to smart layout under the author's control, with rich magazine style layouts for full screen viewing, switching to simpler layouts when the window is shrunk.

The SGML Open consortium is promoting use of DSSSL Lite by James Clark. This is a simplified subset of DSSSL - the document style semantics specification language. DSSSL is a ISO standard for representing presentation semantics for SGML documents, but is much too complex in its entirety to be well suited to the World Wide Web. Håkon Lie maintains a list of pointers to work on style sheets.

Transition Strategy from HTML 2.0

The use of the MIME content type: "text/html; version=3.0" is recommended to prevent existing HTML 2.0 user agents screwing up by attempting to show 3.0 documents. Tests have shown that the suggested content type will safely cause existing user agents to display the save to file dialog rather than incorrectly displaying the document as if it were HTML 2.0.

To make it easy for servers to distinguish 3.0 documents from 2.0 documents, it is suggested that 3.0 files are saved with the extension ".html3" (or ".ht3" for PCs). Servers can also exploit the accept headers in HTTP requests from HTML user agents, to distinguish whether each client can or cannot support HTML 3.0. This makes it practical for information providers to start providing HTML 3.0 versions of existing documents for newer user agents, without impacting older user agents. It is envisaged that programs will be made available for automatic down conversion of 3.0 to 2.0 documents. This conversion could be carried out in batch mode, or on the fly (with caching for greater efficiency).

Design Guidelines

The HTML 3.0 draft specification has been written to the following guidelines.

Lingua Franca for the Web

HTML is intended as a common medium for tying together information from widely different sources. A means to rise above the interoperability problems with existing document formats, and a means to provide a truly open interface to proprietary information systems.


The first version of HTML was designed to be extremely simple, both to author and to write browsers for. This has played a major role in the incredibly rapid growth of the World Wide Web. HTML 3.0 provides a clean superset of HTML 2.0 adding high value features such as tables, text flow around figures and math, while still remaining a simple document format. The pressures to adopt the complexities of traditional SGML applications has been resisted, for example the Department of Defense's CALS table model or the ISO 12083 math DTD.


As time goes by, people's expectations change, and more will be demanded of HTML. One manifestation of this is the pressure to add yet more tags. HTML 3.0 introduces a means for subclassing elements in an open-ended way. This can be used to distinguish the role of a paragraph element as being a couplet in a stansa, or a mathematical term as being a tensor. This ability to make fresh distinctions can be exploited to impart distinct rendering styles or to support richer search mechanisms, without further complicating the HTML document format itself. Scaleability is also achieved via URI based links for embedding information in other formats. Initially limited to a few image formats, inline support is expected to rapidly evolve to cover drawing formats, video, distributed virtual reality and a general means for embedding other applications.

Platform Independence

HTML is designed to allow rendering on a very wide range of devices, from clunky teletypes, to terminals, DOS, Windows, Macs and high end Workstations, as well as non-visual media such as speech and braille. In this, it allows users to exploit the legacy of older equipment as well as the latest and best of new machines. HTML 3.0 provides for improved support for non-graphical clients, allowing for rich markup in place of the figures shown on graphical clients. HTML can be rendered on a wide variety of screen sizes, using a scrolling or paged model. The fonts and presentation can be adjusted to suit the resources available in the host machine and the user's preferences.

Content not Presentation Markup

Information providers are used to tight control over the final appearence of documents. The need for platform independence weighs against this, but there is still a strong pressure to find appropriate means for information providers to express their intentions. The experience with proprietary document formats has shown the dangers of mixing presentation markup with content (or structural) markup. It becomes difficult to apply different presentation styles. It becomes painful to incorporate material from different sources (with different presentation styles). It becomes difficult to be truly platform independent. As a result, HTML 3.0 is designed for use with linked style information that defines the intended presentation style for each element. Style sheets can be expressed in a platform independent fashion or used to provide more detailed control for particular classes of clients or output media.

Support for Cascaded Style Sheets

For the Web, it is valuable to allow for a cascading of style preferences. The client has certain built-in preferences; the publisher may require a particular house style, e.g. for brand distinction; the author may feel the need to override the house style for special cases; the end-user may feel strongly about certain things, e.g. large fonts for easier visibility or avoiding certain colors due to an inability to distinguish between them. HTML 3.0 supports style sheets via the use of the LINK element to reference a style sheet with a URI. Authors can place overrides in separate style sheets or include them in the document head within the STYLE element. The effectiveness of caching mechanisms for speeding up the retrieval of style sheets is enhanced by the separation of style information into generic commonly used style sheets, and overrides specific to this document.

Support for Non-Visual Media

HTML 3.0 is designed to cater for the needs of the visually impaired. Markup for inline figures includes support for rich descriptions, along with hypertext links that double up as defining geometric hotzones for graphical browsers, simplifying the author's job in catering for the different groups of users. Table markup includes provision for abbreviated row and column names for each cell, which are essential for conversion to speech or braille. Math markup treats formulae and equations as hierarchies of expressions. This allows disambiguating pauses to be inserted in appropriate places during conversion to speech.

Support for different ways of creating HTML

HTML 3.0 has been designed to be created in a variety of different ways. It is deliberately simple enough to type in by hand. It can be authored using wysiwyg editors for HTML, or it can be generated via export filters from common word processing formats, or other SGML applications.