The HTML 3.2 standard doesn't specify how user agents (the standard's term for a browser) should break up long lines of text to make them fit into the available display width. In practice, nearly all browsers use "word wrapping" to break up text lines; this means that the browser inserts line breaks at word boundaries only, so that an individual word is never split across two lines of text. This is the way HTML TADS works by default, too, but HTML TADS provides some additional features that let the game override the default rules and take control of how text is wrapped. This section describes these advanced line breaking controls.
To start with, HTML TADS has two basic modes for text wrapping: word and character.
In word-wrapping mode, the formatter will only break a line between words. This is the default mode, and is the style that almost always applies to languages like English that use groups of letters to represent words. The formatter's rules for determining word boundaries are simple: a word boundary is a space, or a hyphen (but a break can occur only to the right of a hyphen, and only when the hyphen isn't followed by another hyphen).
In character-wrapping mode, the formatter can break a line anywhere. This mode is especially applicable to languages like Chinese in which each character represents an entire word. In Chinese, convention allows line breaks to occur almost anywhere; each glyph is a separate word, so there is no need to keep most pairs of adjacent glyphs together on one line.
The game sets the wrapping mode using the <WRAP> tag (which is a TADS extension, not a standard HTML tag). <WRAP WORD> sets word-wrapping mode, and <WRAP CHAR> sets character-wrapping mode. The interpreter always starts in word-wrapping mode. This tag is not a container, but simply an in-line mode switch, so there is no </WRAP> tag; to change out of the current mode, simply use another <WRAP> tag with the new mode.
The default line-breaking rules -- in both word-wrap and character-wrap modes -- are very simple, and aren't adequate for every situation. For example, a game written in English might have some words that include hyphens that should always be kept together on one line, or a word that includes some other embedded punctuation that could serve as line-break points if needed. As another example, a game written in Chinese wouldn't really want line breaks to occur just anywhere; the conventions for line-breaking in Chinese actually require that certain sequences of characters, such as certain groups of punctuation marks, be kept together, and don't allow line breaks to occur just before or after certain types of punctuation.
It would be impossible to anticipate every possible line-breaking rule for every language and every game, so HTML TADS doesn't even try; instead, the interpreter uses the rather simple set of rules outlined above by default, but provides a set of special control sequences that allow the game to override the default behavior whenever it wants. These special controls are described below.
The first special control prevents a line break, and is called the "zero-width non-breaking space." This is a special markup, written as "&zwnbsp;". It's a "zero-width" character, which means that it doesn't show up on the display: it's simply invisible as far as the user is concerned. The "non-breaking" part is the special feature: this tells the interpreter that it cannot break the line here, even if it otherwise could.
To be more specific, the zero-width non-breaking space prevents a line break from occurring between the two characters adjacent to it. Essentially, this control is a bit of glue that sticks so strongly to the characters on both sides that the line breaking rules can't tear them apart.
Note that the glue only sticks to one side of each adjacent character. So, if you put a &zwnbsp character just before a space, and you're in word-wrapping mode, the formatter can still break the line after the space. If you want to prevent the space from being a line break at all, you have to put a &zwnbsp character on both sides of the space-one before and one after.
The second special control is the "zero-width space," and it does essentially the opposite of the zero-width non-breaking space: this character enables a line break where the rules would otherwise not allow it. This markup is written "&zwsp;". Like &zwnbsp, this is a zero-width character, so it's invisible on the display. Otherwise, though, it counts as a space -- which means that the formatter, even in word-wrap mode, is free to break the line between the two adjacent characters, as though they were separated by an ordinary space.
You can use a zero-width space to add your own rules about where lines can be broken. For example, suppose your game has a bunch of words where you're using an equals sign as though it were a hyphen, and you want the formatter to be able to break to the right of these equals signs just like it would for hyphens. To do this, you would simply insert a &zwsp character immediately after each of these equals signs; this would tell the interpreter that it can break the line just after the hyphen if necessary.
Standard HTML defines its own non-breaking space character, written as " ". This character is displayed exactly like an ordinary space (the kind you get by pressing the space bar on the keyboard), but it behaves as though it were a non-space character: the formatter never breaks a line at a non-breaking space (thus the name), it never combines a non-breaking space with adjacent ordinary spaces, and it never trims a non-breaking space from the beginning or end of a line. For line-breaking purposes, the non-breaking space behaves as though it were an alphabetic character.
Apart from the obvious difference in visual size, the non-breaking space differs from the zero-width non-breaking space in that the zero-width version is like glue -- it prevents a line break from being inserted between the adjacent characters. The ordinary non-breaking space, however, merely acts like any other non-space character; if the line-breaking rules allow it, the formatter can break the line immediately before or after an ordinary non-breaking space.
Standard HTML defines another special control, the "soft hyphen," written as "­". This character tells the interpreter that it can break a word with hyphenation at a particular point, but that it doesn't have to.
Soft hyphens are normally invisible, so you can freely insert them into words without adding visual clutter. When the formatter decides to take advantage of a soft hyphen to break a line, though, the soft hyphen is displayed as a normal hyphen at the end of the line.