Contents:
Grammatical Conventions
The Grammar
For the most part, the exact syntax of an HTML document is not rigidly enforced by a browser. This gives authors wide latitude in creating documents and gives rise to documents that work on most browsers, but are actually incompatible with the standard. Stick to the standards unless your documents are fly-by-night affairs.
The HTML standard explicitly defines the ordering and nesting of tags and document elements. This syntax is embedded within the HTML Document Type Definition and is not readily understood by those not versed in SGML (see Appendix C, The HTML DTD). Accordingly, we provide an alternate definition of the allowable HTML syntax, using a fairly common tool called a ``grammar.''
Grammar, whether it defines English sentences or HTML documents, is just a set of rules that indicate the order of language elements. These language elements can be divided into two sets: terminal--the actual words of the language--and nonterminal--all other grammatical rules. In HTML, the words correspond to the embedded markup tags and text in a document.
To use the grammar to create a valid HTML document, follow the order of the rules to see where the tags and text may be placed to create a valid HTML document.
We use a number of typographic and punctuation conventions to make our grammar easy to understand.
For our HTML grammar, we denote the terminals with a bold, monospaced Courier typeface. The nonterminals appear in italicized text.
We also use a simple naming convention for the majority of our nonterminals: if one defines the syntax of a specific HTML tag, its name will be the tag name followed by ``_tag.'' If a nonterminal defines the various language elements that may be nested within a certain tag, its name will be the tag name followed by ``_content.''
For example, if you are wondering exactly which elements are allowed within an <a> tag, you can look for the a_content rule within the grammar. Similarly, to determine the correct syntax of a definition list created with the <dl> tag, look for the dl_tag rule.
Each rule in the grammar starts with the rule's name, followed by the replacement symbol (::=) and the rule's value. We've intentionally kept the grammar simple, but we do use three punctuation elements to denote alternation, repetition, and optional elements in the grammar.
Alternation indicates that a rule may actually have several different values, and you must choose exactly one of them. Vertical bars (|) separate the alternatives for the rule.
For example, the heading rule is equivalent to any one of six HTML heading tags, and so appears in the table as:
heading | ::= | h1_tag |
| | h2_tag | |
| | h3_tag | |
| | h4_tag | |
| | h5_tag | |
| | h6_tag |
The heading rule tells us that wherever the heading nonterminal appears in a rule, you can replace it with exactly one of the actual heading tags.
Repetition indicates that an element within a rule may be repeated some number of times. Repeated elements are enclosed in curly braces ({...}). The closing brace has a subscripted number other than one if the element must be repeated a minimum number of times.
For example, the <ul> tag may only contain <li> tags, or it may actually be empty. The rule, therefore, is:
ul_tag | ::= | <ul> |
{li_tag }_0 | ||
</ul> |
The rule says that the syntax of the <ul> tag requires the <ul> tag, zero or more <li> tags, followed by a closing </ul> tag.
We spread this rule across several lines and indented some of the elements to make it more readable only; it does not imply that your documents must actually be formatted this way.
Some elements may appear in a document, but are not required. Optional elements are enclosed in square brackets ([ and ]).
The <table> tag, for example, has an optional caption:
table_tag | ::= | <table> |
[ caption_tag ] | ||
{tr_tag }_0 | ||
</table> |
In addition, the rule says that a table begins with the <table> tag, followed by an optional caption, zero or more table-row tags, and ends with the </table> tag.
Our grammar stops at the tag level; it does not delve further to show the syntax of each tag, including tag attributes. For these details, refer to the HTML Quick Reference card included with this book.
The HTML standard defines a few specific kinds of content that correspond to various types of text. We use these content types throughout the grammar. They are:
Text is interpreted exactly as specified; no character entities or style tags are recognized.
Regular characters in the document character encoding, along with character entities denoted by the ampersand character.
Like plain_text, with physical- and content-based style tags allowed.