Whitespace Collapsing

A whitespace character is a character data element that represents white space when text is
rendered for display by a computer.

For example, a ''space'' character (, ASCII 32) represents blank space such as a word divider in a Western script.

A printable character results in output when rendered,
but a whitespace character does not.
Instead, whitespace characters define the layout of text to a limited degree, interrupting the normal sequence of rendering characters next to each other.
The output of subsequent characters is typically shifted to the right (or to the left for right-to-left script) or to the start of the next line.
The effect of multiple sequential whitespace characters is cumulative such that the next printable character is rendered at a location based on the accumulated effect of preceding whitespace characters.

The origin of the term ''whitespace'' is rooted in the common practice of rendering text on white paper. Normally, a whitespace character is ''not'' rendered as white. It affects rendering, but it is not itself rendered.

Overview

A space character typically inserts horizontal space that is about as wide as a letter. For a monospaced font the width is the width of a letter, and for a variable-width font the width is font-specific. Some fonts support multiple space characters that have different widths.

A tab character typically inserts horizontal space that is based on tab stops which vary by application.

A newline character sequence typically moves the render output location to the beginning of the next line. If one follows text, it does not actually result in whitespace. But, two sequential newline sequences between text blocks results in a blank line between the blocks. The height of the blank line varies by application.

Using whitespace characters to layout text is a convention. Applications sometimes render whitespace characters as visible markup so that a user can see what is normally not visible.

Typically, a user types a space character by pressing , a tab character by pressing and newline by pressing .

Unicode

The table below lists the twenty-five characters defined as whitespace ("WSpace=Y", "WS") characters in the Unicode Character Database. Seventeen use a definition of whitespace consistent with the algorithm for bidirectional writing ("Bidirectional Character Type=WS") and are known as "Bidi-WS" characters. The remaining characters may also be used, but are not of this "Bidi" type.

''Note: Depending on the browser and fonts used to view the following table, not all spaces may be displayed properly.''

Substitute images

Unicode also provides some visible characters that can be used to represent various whitespace characters, in contexts where a visible symbol must be displayed:

; Exact space
* The Cambridge Z88 provided a special "exact space" (code point 160 aka 0xA0) (invokable by key shortcut ), displayed as "…" by the operating system's display driver. It was therefore also known as "dot space" in conjunction with BBC BASIC.
*Under code point 224 (0xE0) the computer also provided a special three-character-cells-wide SPACE symbol "SPC" (analogous to Unicode's single-cell-wide U+2420).

Non-space blanks

* The Braille Patterns Unicode block contains , a Braille pattern with no dots raised. Some fonts display the character as a fixed-width blank, however the Unicode standard explicitly states that it does not act as a space.Unicode chart U+2800
braille patterns
* Unicode's coverage of the Korean alphabet includes several code points which represent the absence of a written letter, and thus do not display a glyph:
** Unicode includes a Hangul Filler character in the Hangul Compatibility Jamo block (). This is classified as a letter, but displayed as an empty space, like a Hangul block containing no jamo. It is used in KS X 1001 Hangul combining sequences to introduce them or denote the absence of a letter in a position, but not in Unicode's combining jamo system.
** Unicode's combining jamo system uses similar Hangul Choseong Filler and Hangul Jungseong Filler characters to denote the absence of a letter in initial or medial position within a syllable block, which are included in the Hangul Jamo block (, ).
** Additionally, a Halfwidth Hangul Filler is included in the Halfwidth and Fullwidth Forms (), which is used when mapping from encodings which include characters from both Johab (or Wansung) and N-byte Hangul (or its EBCDIC counterpart), such as IBM-933, which includes both Johab and EBCDIC fillers.

Whitespace and digital typography

On-screen display

Text editors, word processors, and desktop publishing software differ in how they represent whitespace on the screen, and how they represent spaces at the ends of lines longer than the screen or column width. In some cases, spaces are shown simply as blank space; in other cases they may be represented by an interpunct or other symbols. Many different characters (described below) could be used to produce spaces, and non-character functions (such as margins and tab settings) can also affect whitespace.

Many of the Unicode space characters were created for compatibility with classic print typography.

Even if digital typography has algorithmic kerning and justification, those space characters can be used to supplement the electronic formatting when needed.

Variable-width general-purpose space

In computer character encodings, there is a normal general-purpose space (Unicode character U+0020) whose width will vary according to the design of the typeface. Typical values range from 1/5 em to 1/3 em (in digital typography an em is equal to the nominal size of the font, so for a 10-point font the space will probably be between 2 and 3.3 points). Sophisticated fonts may have differently sized spaces for bold, italic, and small-caps faces, and often compositors will manually adjust the width of the space depending on the size and prominence of the text.

In addition to this general-purpose space, it is possible to encode a space of a specific width. See the table below for a complete list.

Hair spaces around dashes

Em dashes used as parenthetical dividers, and en dashes when used as word joiners, are usually set continuous with the text. However, such a dash can optionally be surrounded with a hair space, U+200A, or thin space, U+2009. The hair space can be written in HTML by using the numeric character references   or  , or the named entity  . The thin space is named entity   and numeric references   or  . These spaces are much thinner than a normal space (except in a monospaced (non-proportional) font), with the hair space in particular being the thinnest of horizontal whitespace characters.

Computing applications

Programming languages

In most programming language syntax, whitespace characters can be used to separate tokens. For a free-form language, whitespace characters are ignored by code processors (i.e. compiler). Even when language syntax requires white space, often multiple whitespace characters are treated the same as a single. In an off-side rule language, indentation white space is syntactically significant. In the satirical and contrarian language called Whitespace, whitespace characters are the only significant characters and normal text is ignored.

Good use of white space in source code can group related logic and make the code easier to understand.
Excessive use of whitespace, including at the end of a line where it provides no rendering behavior, is considered a nuisance.

Most languages only recognize whitespace characters that have an ASCII code. They disallow most or all of the Unicode codes listed above. The C language defines whitespace characters to be "space, horizontal tab, new-line, vertical tab, and form-feed". The HTTP network protocol requires different types of whitespace to be used in different parts of the protocol, such as: only the space character in the status line, CRLF at the end of a line, and "linear whitespace" in header values.

Command-line parsing

Typical command-line parsers use the space character to delimit arguments. A value with an embedded space character is problematic since it causes the value to parse as multiple arguments. Typically, a parser allows for escaping the normal argument parsing by enclosing the text in quotes.

Consider that one wants to list the files in directory named "foo bar". This command instead lists the files matching either "foo" or "bar":

ls foo bar

This command correctly specifies a single argument:

ls "foo bar"

Markup languages

Some markup languages, such as SGML, preserve whitespace as written.

Web markup languages such as XML and HTML treat whitespace characters specially, including space characters, for programmers' convenience. One or more space characters read by conforming display-time processors of those markup languages are collapsed to 0 or 1 space, depending on their semantic context. For example, double (or more) spaces within text are collapsed to a single space, and spaces which appear on either side of the "=" that separates an attribute name from its value have no effect on the interpretation of the document. Element end tags can contain trailing spaces, and empty-element tags in XML can contain spaces before the "/>". In these languages, unnecessary whitespace increases the file size, and so may slow network transfers. On the other hand, unnecessary whitespace can also inconspicuously mark code, similar to, but less obvious than comments in code. This can be desirable to prove an infringement of license or copyright that was committed by copying and pasting.

In XML attribute values, sequences of whitespace characters are treated as a single space when the document is read by a parser. Whitespace in XML element content is not changed in this way by the parser, but an application receiving information from the parser may choose to apply similar rules to element content. An XML document author can use the xml:space="preserve" attribute on an element to instruct the parser to discourage the downstream application from altering whitespace in that element's content.

In most HTML elements, a sequence of whitespace characters is treated as a single ''inter-word separator'', which may manifest as a single space character when rendering text in a language that normally inserts such space between words. Conforming HTML renderers are required to apply a more literal treatment of whitespace within a few prescribed elements, such as the pre tag and any element for which CSS has been used to apply pre-like whitespace processing. In such elements, space characters will not be "collapsed" into inter-word separators.

In both XML and HTML, the non-breaking space character, along with other non-"standard" spaces, is not treated as collapsible "whitespace", so it is not subject to the rules above.

File names

Such usage is similar to multiword file names written for operating systems and applications that are confused by embedded space codes—such file names instead use an underscore (_) as a word separator, as_in_this_phrase.

Another such symbol was . This was used in the early years of computer programming when writing on coding forms. Keypunch operators immediately recognized the symbol as an "explicit space". It was used in BCDIC, EBCDIC, and ASCII-1963.

See also

* Carriage return
* Em (typography)
* En (typography)
* Form feed
* Indent style
* Line feed
* Newline
* Programming style
* Prosigns for Morse code
* for the white-space character class.
* Space bar
* Space (punctuation)
* Tab key
* Trimming (computer programming)
* Whitespace (programming language)
* Zero-width space

References

External links

Property List of Unicode Character Database

{{DEFAULTSORT:Whitespace (Computer Science)
Character encoding
Source code