Overview.src.html

<h1>CSS Ruby Layout Module Level 1</h1>

<pre class='metadata'>
Shortname: css-ruby
Level: 1
Status: ED
Group: csswg
ED: http://dev.w3.org/csswg/css-ruby-1/
TR: http://www.w3.org/TR/css-ruby-1/
Previous Version: http://www.w3.org/TR/2013/WD-css3-ruby-20130919/
Editor: Elika J. Etemad, Invited Expert, http://fantasai.inkedblade.net/contact
Editor: Koji Ishii, Invited Expert, kojiishi@gluesoft.co.jp
Abstract: “Ruby” are short runs of text alongside the base text,
Abstract: typically used in East Asian documents to indicate pronunciation or to provide a short annotation.
Abstract: This module describes the rendering model and formatting controls related to displaying ruby annotations in CSS.
Ignored terms: collapsible, collapsible white space, white space, segment break, soft wrap opportunity, justification opportunity, justification opportunities, characters
</pre>

<!--

Issues:
	bidi
	box layout/sizing
	clean up inter-character vs. parallel layout requirements

Redo all examples with consistent font. (M+ 2p?)

-->

<h2 id="intro">
Introduction</h2>

	<p><em>This section is not normative.</em>

	<div class="issue">
		<p>Depending on how implementers feel about it,
		we might want to limit (via the at-risk list) CSS Ruby Layout Level 1, like HTML5,
		to two levels of annotation.
		Subsequent levels would be inlined immediately after their ruby segment.
		(Not sure how to represent that limitation in <code>@supports</code>, though.)

		<p>FWIW, the main complexity item of concern with >2 levels is nested ruby.
		The next most complex feature, after nested ruby,
		is allowing two levels of ruby on the same side of the text.
		Removing either of these features, however, boxes us into a corner.
		By removing nested ruby handling,
		we'd create a dependency on nested ruby not being handled optimally
		(unless we did something dumb and always obviously broken somehow).
		By forcing a second annotation to be on the opposite side of the first,
		we create an assumption that it can never be placed on the same side,
		and that 'ruby-position' rules that would place the second annotation
		on the same side as the first have no effect.
	</div>

<h3 id="placement">
Module interactions</h3>

	<p>This module extends the inline box model of CSS Level 2 [[!CSS21]]
	to support ruby.

	<p>None of the properties in this module apply to the <code>::first-line</code> or
	<code>::first-letter</code> pseudo-elements.

<h3 id="values">
Values</h3>

	<p>This specification follows the
	<a href="http://www.w3.org/TR/CSS21/about.html#property-defs">CSS property
	definition conventions</a> from [[!CSS21]]. Value types not defined in
	this specification are defined in CSS Level 2 Revision 1 [[!CSS21]].
	Other CSS modules may expand the definitions of these value types: for
	example [[CSS3VAL]], when combined with this module, expands the
	definition of the <var>&lt;length&gt;</var> value type as used in this specification.</p>

	<p>In addition to the property-specific values listed in their definitions,
	all properties defined in this specification also accept the
	<a href="http://www.w3.org/TR/CSS21/cascade.html#value-def-inherit">inherit</a>
	keyword as their property value. For readability it has not been repeated
	explicitly.

<h3 id="diagram-conventions">
Diagram conventions</h3>

	<p>Many typographical conventions in East Asian typography depend
	on whether the character rendered is wide (CJK) or narrow (non-CJK).
	There are a number of illustrations in this document
	for which the following legend is used:

	<dl>
		<dt><img alt="Symbolic wide-cell glyph representation" width="39" height="39" src="images/fullwidth.gif">
		<dd>Wide-cell glyph (e.g. Han) that is the <var>n</var>th character in the text run.
		They are typically sized to 50% when used as annotations.
		<dt><img alt="Symbolic narrow-cell glyph representation" width="19" height="39" src="images/halfwidth.gif">
		<dd>Narrow-cell glyph (e.g. Roman) which is the <var>n</var>th glyph in the text run.
	</dl>

	<p>The orientation which the above symbols assume in the diagrams
	corresponds to the orientation that the glyphs they represent
	are intended to assume when rendered by the user agent.
	Spacing between these characters in the diagrams is incidental,
	unless intentionally changed to make a point.

<h3 id="ruby-def">
What is ruby?</h3>

	<p><dfn>Ruby</dfn> is the commonly-used name for a run of text
	that appears alongside another run of text (referred to as the “base”)
	and serves as an annotation or a pronunciation guide associated with that run of text.

	<p>The following figures show two examples of Ruby,
	a simple case and one with more complicated structure.

	<div class="example">
		<p>In this first example, a single annotation is used to annotate the base text.
		<div class="figure">
			<p><img src="images/licence.png"
			        alt="Example of ruby applied on top of a Japanese expression">
			<p class="caption">Example of ruby used in Japanese (simple case)
		</div>
		<p>In Japanese typography, this case is sometimes called
		<span lang="ja">taigo</span> ruby or group-ruby (per-word ruby),
		because the annotation as a whole is associated
		with multi-character word (as a whole).
	</div>

	<div class="example">
		<p>In this second example,
		two levels of annotations are attached to a base sequence:
		the hiragana characters on top refer to the pronunciation of each of the base kanji characters,
		while the words “Keio” and “University” on the bottom are annotations describing the English translation.
		<div class="figure">
			<p><img src="images/ruby-univ.gif"
			        alt="Example showing complex ruby with annotation text over and under the base characters">
			<p class="caption">Complex ruby with annotation text over and under the base characters
		</div>
		<p>
		<p>Notice that to allow correct association between the hiragana characters and 
		their corresponding Kanji base characters,
		the spacing between these Kanji characters is adjusted.
		(This happens around the fourth Kanji character in the figure above.)
		To avoid variable spacing between the Kanji characters in the example above
		the hiragana annotations can be styled as a <i>collapsed annotation</i>,
		which will look more like the group-ruby example earlier.
		However because the base-annotation pairings are recorded in the ruby structure,
		if the text breaks across lines, the annotation characters will stay
		correctly paired with their respective base characters.
	</div>

	<p><i>Ruby</i> formatting as used in Japanese is described in JIS X-4051 [[JIS4051]] (in Japanese)
	and in Requirements for Japanese Text Layout [[JLREQ]] (in English and Japanese)].
	In HTML, ruby structure and markup to represent it is described
	in the Ruby Markup Extension specification.
	This module describes the CSS rendering model
	and formatting controls relevant to ruby layout of such markup.

<h2 id="ruby-model">
Ruby Formatting Model</h2>

	<p>The CSS ruby model is based on
	the <a href="http://darobin.github.io/html-ruby/">HTML Ruby Markup Extension</a>
	and <a href="http://www.w3.org/TR/ruby/">XHTML Ruby Annotation Recommendation</a> [[RUBY]].
	In this model, a ruby structure consists of
	one or more <i>ruby base</i> elements representing the base (annotated) text,
	associated with one or more levels of <i>ruby annotation</i> elements representing the annotations.
	The structure of ruby is similar to that of a table:
	there are “rows” (the <i title="base level">base text level</i>, each <i>annotation level</i>)
	and “columns” (each <i>ruby base</i> and its corresponding <i>ruby annotations</i>).

	<p>Consecutive bases and annotations are grouped together into <i>ruby segments</i>.
	Within a <i>ruby segment</i>, a <i>ruby annotation</i> may span multiple <i>ruby bases</i>.

	<p class="note">In HTML, a single <code>&lt;ruby&gt;</code> element may contain multiple <i>ruby segments</i>.
	(In the XHTML Ruby model, a single <code>&lt;ruby&gt;</code> element can only contain one <i>ruby segment</i>.)

<h3 id="ruby-display">
Ruby-specific 'display' property values</h3>

	<p>For document languages (such as XML applications) that do not have pre-defined ruby elements,
	authors must map document language elements to ruby elements;
	this is done with the 'display' property.

	<table class="propdef">
		<tr>
			<th>Name:
			<td>display
		<tr>
			<th><a href="#values">New Values</a>:
			<td>ruby | ruby-base | ruby-text | ruby-base-container | ruby-text-container
	</table>

	<p>The following new 'display' values assign ruby layout roles to an arbitrary element:

	<dl dfn-for=display>
		<dt>''ruby''
			<dd>Specifies that an element generates a <dfn title="ruby container | ruby container box">ruby container box</dfn>.
			(Corresponds to HTML/XHTML <code>&lt;ruby&gt;</code> elements.)
		<dt>''ruby-base''
			<dd>Specifies that an element generates a <dfn title="ruby base box | ruby base">ruby base box</dfn>.
			(Corresponds to HTML/XHTML <code>&lt;rb&gt;</code> elements.)
		<dt>''ruby-text''
			<dd>Specifies that an element generates a <dfn title="ruby annotation box | ruby annotation | annotation">ruby annotation box</dfn>.
			(Corresponds to HTML/XHTML <code>&lt;rt&gt;</code> elements.)
		<dt>''ruby-base-container''
			<dd>Specifies that an element generates a <dfn title="ruby base container box | ruby base container">ruby base container box</dfn>.
			(Corresponds to XHTML <code>&lt;rbc&gt;</code> elements; generated as an anonymous box in HTML.)
		<dt>''ruby-text-container''
			<dd>Specifies that an element generates a <dfn title="ruby annotation container box | ruby annotation container">ruby annotation container box</dfn>.
			(Corresponds to HTML/XHTML <code>&lt;rtc&gt;</code> elements.)
	</dl>

<h3 id="box-fixup">
Anonymous Ruby Box Generation</h3>

	<p>The CSS model does not require that the document language
	include elements that correspond to each of these components.
	Missing parts of the structure are implied through the anonymous box generation rules
	<a href="http://www.w3.org/TR/CSS21/tables.html#anonymous-boxes">similar to those used to normalize tables</a>. [[!CSS21]]

	<ol>
		<li id="anon-gen-inlinize"><strong>Inlinize block-level boxes:</strong>
			Any in-flow block-level boxes directly contained by a
			<i>ruby container</i>,
			<i>ruby base container</i>,
			<i>ruby annotation container</i>,
			<i>ruby base box</i>,
			or <i>ruby annotation box</i>
			are forced to be inline-level boxes,
			and their 'display' value computed accordingly.
			For example,
			the 'display' property of an in-flow element with ''display: block''
			parented by an element with ''display: ruby-text''
			computes to ''inline-block''.
			This computation occurs after any intermediary anonymous-box fixup
			(such as that required by internal table elements).

		<li id="anony-gen-trim-space"><strong>Trim leading/trailing white space:</strong>
			Any white space within a ruby structure is discarded
			at the beginning and end of
			a <i>ruby container</i>, <i>ruby annotation container</i>, or <i>ruby base container</i>.

		<li id="anon-gen-bare-inlines"><strong>Wrap misparented inline-level content:</strong>
			Any consecutive sequence of text and inline-level boxes
			directly parented by a <i>ruby container</i> or <i>ruby base container</i>
			is wrapped in an anonymous <i>ruby base</i>.
			Similarly, any consecutive sequence of text and inline-level boxes
			directly parented by a <i>ruby annotation container</i>
			is wrapped in an anonymous <i>ruby annotation</i>.

			However, if the anonymous box so constructed contains only <i>white space</i>,
			it is considered <dfn>intra-ruby white space</dfn>
			and is either discarded as <i>inter-level white space</i>
			or preserved as <i>intra-level white space</i>
			as described below.

		<li id="anon-gen-discard-space"><strong>Remove inter-level white space:</strong>
			Any <i>intra-ruby white space</i>
			whose immediately adjacent siblings match one of the patterns below
			is <dfn>inter-level white space</dfn>
			and is removed, as if it had ''display: none''.
			<table class="data">
			<thead>
				<tr><th>Previous box
				    <th>Next box
			<tbody>
				<tr><td><i>ruby base</i> or <i>ruby base container</i>
				    <td><i>ruby annotation</i> or <i>ruby annotation container</i>
				<tr><td><i>ruby annotation container</i>
				    <td><i>ruby annotation container</i>
				<tr><td><i>ruby annotation</i>
				    <td><i>ruby annotation container</i>
				<tr><td><i>ruby annotation container</i>
				    <td><i>ruby annotation</i>
			</table>

		<li id="anon-gen-interpret-space"><strong>Interpret intra-level white space:</strong>
			Any <i>intra-ruby white space</i>
			whose immediately adjacent siblings match one of the patterns below
			is assigned the box type and subtype defined in the table below:
			<table class="data">
			<colgroup span=2></colgroup>
			<colgroup span=2></colgroup>
			<thead>
				<tr><th>Previous box
				    <th>Next box
				    <th>Box type
				    <th>Subtype
			<tbody>
				<tr><td><i>ruby base</i>
				    <td><i>ruby base</i>
				    <td><i>ruby base</i>
				    <td><dfn>inter-base white space</dfn>
				<tr><td><i>ruby annotation</i>
				    <td><i>ruby annotation</i>
				    <td><i>ruby annotation</i>
				    <td><dfn>inter-annotation white space</dfn>
				<tr><td><i>ruby annotation</i> or <i>ruby annotation container</i>
				    <td><i>ruby base</i> or <i>ruby base container</i>
				    <td><i>ruby base</i>
				    <td><dfn>inter-segment white space</dfn>
			</table>
			The <dfn>intra-level white space</dfn> boxes defined above are
			treated specially for pairing and layout. See below.

		<li id="anon-gen-anon-containers"><strong>Generate anonymous level containers:</strong>
			Any consecutive sequence of <i>ruby bases</i> and <i>inter-base white space</i>
			(and not <i>inter-segment white space</i>)
			not parented by a <i>ruby base container</i>
			is wrapped in an anonymous <i>ruby base container</i>.
			Similarly, any consecutive sequence of <i>ruby annotations</i> and <i>inter-annotation white space</i>
			not parented by a <i>ruby annotation container</i>
			is wrapped in an anonymous <i>ruby annotation container</i>.

		<li id="anon-gen-anon-ruby"><strong>Generate anonymous ruby containers</strong>:
			A sequence of <i>ruby base containers</i>, <i>ruby annotation containers</i>, and/or <i>inter-segment white space</i>
			not parented by a <i>ruby container</i>
			is wrapped in an anonymous <i>ruby container</i>.

		<li id="anon-gen-unbreak"><strong>Suppress line breaks:</strong>
			Convert all forced line breaks inside <i>ruby annotations</i> (regardless of 'white-space' value)
			as defined for <i>collapsible</i> segment breaks in <a href="http://www.w3.org/TR/css-text-3/#line-break-transform">CSS Text Level 3 &sect; 4.1.2</a>.
			<p class="issue">The goal of this is to simplify the layout model by suppressing any line breaks within ruby annotations.
			Alternatively we could try to define some kind of acceptable behavior for them.
	</ol>

	<p>For the purpose of the rules above,
	a <i>ruby container</i> directly parented by a <i>ruby container</i>
	is considered to be a <i>ruby base container</i>.
	See <i>nested ruby</i>, below.

	<p class="issue">Make this into an example:
		<img src="http://lists.w3.org/Archives/Public/www-archive/2014Jun/att-0027/PastedGraphic-1.png" style="display: block; width: 100%">

	<p>Once all ruby layout structures are properly parented,
	the UA can start to associate bases with their annotations.

	<p class="note">
	Note that the UA is not required to create any of these anonymous boxes
	(or the anonymous empty <i>intra-level white space</i> boxes in the pairing section)
	in its internal structures,
	as long as pairing and layout behaves as if they existed.

<h4 id="block-ruby">
Block-level Ruby</h4>

	<p>If an element has a computed 'display-inside' of ''ruby''
	and a computed 'display-outside' other than ''inline-level'',
	then it generates two boxes:
	a principal block container box of the required 'display-outside' type,
	and an inline-level <i>ruby container</i>.
	All properties specified on the element apply to the principal box
	(and if inheritable, inherit to the <i>ruby container box</i>).
	This allows styling the element as a block,
	while correctly maintaining the internal ruby structure.

	<p>See the <a href="http://www.w3.org/TR/css-display/">CSS Display Module</a>
	for more information on 'display-inside' and 'display-outside'.

<h3 id="ruby-pairing">
Ruby Pairing and Annotation Levels</h3>

	<p>Annotation <dfn>pairing</dfn> is the process of associating
	<i>ruby annotations</i> with <i>ruby bases</i>.
	Each <i>ruby annotation</i> is associated with one or more <i>ruby bases</i>,
	and is said to <dfn>span</dfn> those bases.
	(A <i>ruby annotation</i> that <i>spans</i> multiple bases is called
	a <dfn>spanning annotation</dfn>.)

	A <i>ruby base</i> is can be associated with
	only one <i>ruby annotation</i> per <i>annotation level</i>.
	However, if there are multiple <i>annotation levels</i>,
	it can be associated with multiple <i>ruby annotations</i>.

	Once pairing is complete, ruby “column” units are defined,
	each represented by a single <i>ruby base</i>
	and one <i>ruby annotation</i> (possibly an empty, anonymous one)
	from each <i>annotation level</i> in its <i>ruby segment</i>.


<h4 id="segment-pairing">
Segment Pairing</h4>

	<p>A ruby structure is divided into <dfn>ruby segments</dfn>,
	each consisting of a single <i>ruby base container</i>
	followed by one or more <i>ruby annotation containers</i>.
	Each <i>ruby annotation container</i> in a <i>ruby segment</i>
	represents one <dfn title="annotation level | level">level</dfn> of annotation for the base text:
	the first one represents the first level of annotation,
	the second one represents the second level of annotation,
	and so on.
	The <i>ruby base container</i> represents the <dfn>base level</dfn>.
	The <i>ruby base container</i> in each segment is thus paired
	with each of the <i>ruby annotation containers</i> in that segment.

	<p>In order to handle degenerate cases, some empty anonymous containers are assumed:
	<ul>
		<li>If the first child of a <i>ruby container</i> is a <i>ruby annotation container</i>,
		an anonymous, empty <i>ruby base container</i> is assumed to exist before it.
		<li>Similarly, if the <i>ruby container</i> contains consecutive <i>ruby base containers</i>,
		anonymous, empty <i>ruby annotation containers</i> are assumed to exist between them.
	</ul>

	<p><i>Inter-segment white space</i> is effectively a <i>ruby segment</i> of its own.

<h4 id="base-annotation-pairing">
Unit Pairing</h4>

	<p>Within a <i>ruby segment</i>,
	each <i>ruby base</i> in the <i>ruby base container</i>
	is paired with one <i>ruby annotation</i>
	from each <i>ruby annotation container</i> in its <i>ruby segment</i>.
	If there are not enough <i>ruby annotations</i> in a <i>ruby annotation container</i>,
	the last one is paired with any excess <i>ruby bases</i>.
	(If there are not any in the <i>ruby annotation container</i>,
	 an anonymous empty one is assumed to exist.)
	If there are not enough <i>ruby bases</i>,
	any remaining <i>ruby annotations</i> pair with empty, anonymous bases
	inserted at the end of the <i>ruby base container</i>.

	<p>If an implementation supports ruby markup with explicit spanning
	(e.g. XHTML Complex Ruby Annotations),
	it must adjust the pairing rules to pair <i>spanning annotations</i>
	to their bases appropriately.

	<p><i>Intra-level white space</i> does not participate in standard annotation <i>pairing</i>.
	However, if the immediately-adjacent <i>ruby bases</i> or <i>ruby annotations</i>
	are paired
	<ul>
		<li>with two <i>ruby bases</i> or <i title="ruby annotations">annotations</i>
		that surround corresponding <i>intra-level white space</i> in another level,
		then the so-corresponding <i>intra-level white space</i> boxes are also paired.
		<li>with a single spanning <i>ruby annotation</i>,
		then the <i>intra-level white space</i> is also paired to that <i>ruby annotation</i>
		<li>with two <i>ruby bases</i> or <i title="ruby annotations">annotations</i>
		with no intervening <i>intra-level white space</i>,
		then the <i>intra-level white space</i> box pairs with
		an anonymous empty <i>intra-level white space</i> box assumed to exist between them.
	</ul>

	<p class="issue">Insert diagram</p>

<h4 id="nested-pairing">
Complex Spanning and Nested Ruby</h4>

	<p><dfn title="nested ruby"></dfn>When <i>ruby containers</i> are nested,
	pairing begins with the deepest <i>ruby container</i>,
	then expands out.
	From the pairing perspective of the outer <i>ruby container</i>,
	each <i>ruby container</i> nested within another <i>ruby container</i>
	counts as representing a single <i>ruby base</i>/<i>annotation</i> per level.
	The outer <i>ruby container</i>’s <i>ruby annotations</i> paired to the <i>nested ruby</i>
	are therefore paired with (and <i>span</i>) all of the nested <i>ruby container</i>’s <i>ruby bases</i>.
	Each <i>ruby annotation container</i> in the nested <i>ruby container</i>
	occupies the same <i>annotation level</i> in the outer <i>ruby container</i>
	as it does in the inner one
	and participates in its layout as if it were directly contained in the outer <i>ruby container</i>.

	<p>This process is recursive.
	Thus, using nested <i>ruby containers</i> allows the representation
	of complex spanning relationships.

	<p class="issue">This has to be Level 1 because HTML5 allows it, so we have to handle it. Yay HTML5.
	(Also, if we're ever going to support it, it has to be designed into level 1
	or authors will depend on whatever degenerate behavior falls out.)

<h3 id="autohide">
Autohiding Annotations</h3>

	<p>If a <i>ruby annotation</i> has the exact same text content as its base,
	it is <dfn title="hidden ruby annotation | hidden annotation">hidden</dfn>.
	Hiding a <i>ruby annotation</i> does not affect annotation pairing
	or the block-axis positioning of boxes in other <i>levels</i>.
	However the <i>hidden annotation</i> is not visible,
	and it has no impact on layout
	other than to separate adjacent sequences of <i>ruby annotation boxes</i> within its level,
	as if they belonged to separate segments
	and the <i>hidden annotation</i>’s base were not a <i>ruby base</i> but an intervening inline.

	<div class="example">
		<p>This is to allow correct inlined display of annotations
		for Japanese words that are a mix of kanji and hiragana.
		For example, the word <span lang=ja>振り仮名</span> should be inlined as
		<p class="figure">振り仮名(ふりがな)
		<p>and therefore marked up as
		<pre>
<!--		-->&lt;ruby>
<!--		-->  &lt;rb>振&lt;/rb>&lt;rb>り&lt;/rb>&lt;rb>仮&lt;/rb>&lt;rb>名&lt;/rb>
<!--		-->  &lt;rp>(&lt;/rp>&lt;rt>ふ&lt;/rt>&lt;rt>り&lt;/rt>&lt;rt>が&lt;/rt>&lt;rt>な&lt;/rt>&lt;rp>)&lt;/rp>
<!--		-->&lt;ruby></pre>
		<p>However, when displayed as ruby, the “り” should be hidden
		<div class="figure">
			<p><img src="images/furigana-separate.png"
			        alt="Hiragana annotations for 振り仮名 appear, each pronunciation above its kanji base character.">
			<p class="caption">Hiragana ruby for 振り仮名. Notice there is no hiragana annotation above り,  since it is already in hiragana.
		</div>
	</div>

	<p class="note">
		Future levels of CSS Ruby may add controls for this,
		but in this level it is always forced.

	<p>The content comparison for this auto-hiding behavior
	takes place prior to white space collapsing ('white-space') and text transformation ('text-transform')
	and ignores elements (considers only the <code>textContent</code> of the boxes).

<h3 id="white-space">
White Space</h3>

	<p><i>Collapsible white space</i> within a ruby structure is discarded
	<ul>
		<li>at the beginning and end of a <i>ruby container</i>, <i>ruby annotation container</i>, or <i>ruby base container</i>,
		<li>at the beginning/end of a <i>ruby annotation box</i> or <i>ruby base box</i> if white space is not its only contents,
		<li>between a <i>ruby base container</i> and its following <i>ruby annotation container</i>,
		<li>between <i>ruby annotation containers</i>.
	</ul>

	<p>Between <i>ruby segments</i>, between <i>ruby bases</i>, and between <i>ruby annotations</i>, however,
	white space is not discarded.

	<p>Where undiscarded white space is <i>collapsible</i>, it will collapse
	following the standard <a href="http://www.w3.org/TR/css3-text/#white-space-rules">white space processing rules</a>. [[!CSS3TEXT]]
	For <i>collapsible</i> white space between <i>ruby segments</i>, however,
	the contextual text for determining collapsing behavior is given by the <i>ruby bases</i> on either side,
	not the text on either side of the white space in the source document.

	<div class="note">
		<p>Note that the white space processing rules
		cause a white space sequence containing a <i>segment break</i> (such as a line feed)
		to <a href="http://www.w3.org/TR/css3-text/#line-break-transform">collapse to nothing</a> between CJK characters.
		This means that CJK ruby can safely use white space for indentation of the ruby markup.
		For example, the following markup will display without any spaces:
		<pre>
<!--		-->&lt;ruby>
<!--		-->  &lt;rb>東&lt;/rb>&lt;rb>京&lt;/rb>
<!--		-->  &lt;rt>とう&lt;/rt>&lt;rt>きょう&lt;/rt>
<!--		-->&lt;/ruby></pre>
		<p>However, white space that does not contain a <i>segment break</i> does not collapse completely away,
		so this markup will display with a space between the first and second ruby pairs:
		<pre>
<!--		-->&lt;ruby>
<!--		-->  &lt;rb>東&lt;/rb>	&lt;rb>京&lt;/rb>
<!--		-->  &lt;rt>とう&lt;/rt>	&lt;rt>きょう&lt;/rt>
<!--		-->&lt;/ruby></pre>
	</div>

	<p>Any preserved white space is then wrapped in an anonymous box belonging to
	the <i>ruby base container</i> (if between <i>ruby bases</i>),
	<i>ruby annotation container</i> (if between <i>ruby annotations</i>),
	or <i>ruby container</i> (if between <i>ruby segments</i>).
	In the latter case, the text is considered part of the <i>base level</i>.
	Such anonymous boxes do not take part in pairing.
	They merely ensure separation between adjacent bases/annotations.

	<p class="issue">Specify how these anonymous white space boxes impact layout.

	<div class="example">
		<p>These rules allow ruby to be used with space-separated scripts such as Latin.
		For example,
		<pre>
<!--		-->&lt;ruby>
<!--		-->  &lt;rb>W&lt;/rb>&lt;rb>W&lt;/rb>&lt;rb>W&lt;/rb>
<!--		-->  &lt;rt>World&lt;/rt> &lt;rt>Wide&lt;/rt> &lt;rt>Web&lt;/rt>
<!--		-->&lt;/ruby></pre>
		<p>They also ensure that annotated white space is preserved. For example,
		<pre>
<!--		-->&lt;ruby>
<!--		-->  &lt;rb>Aerith&lt;/rb>&lt;rb> &lt;/rb>&lt;rb>Gainsboro&lt;/rb>
<!--		-->  &lt;rt>エアリス&lt;/rt>&lt;rt>・&lt;/rt>&lt;rt>ゲインズブール&lt;/rt>
<!--		-->&lt;/ruby></pre>
	</div>

<h3 id="ruby-layout">
Ruby layout</h3>

	<p>When a ruby structure is laid out,
	its <i>base level</i> is laid out on the line,
	aligned according to its 'vertical-align' property
	exactly as if its <i>ruby bases</i> were a regular sequence of inline boxes.
	Each <i>ruby base container</i> is sized and positioned
	to contain exactly the full height of its <i>ruby bases</i>.

	<p><i>Ruby annotations</i> associated with the <i>base level</i>
	are then positioned with respect to their <i>ruby base boxes</i>
	according to the applicable 'ruby-position' values.
	<i>Ruby annotations</i> within a level (within a single <i>ruby container</i>)
	are aligned to each other as if they were inline boxes
	participating in the same inline formatting context.
	Each <i>ruby annotation container</i> is sized and positioned
	to contain exactly the full height of its <i>ruby annotations</i>.

	<p>A ruby container (or fragment thereof)
	measures as wide as the content of its widest level.
	Similarly, <i>ruby base boxes</i> and <i>ruby annotation boxes</i>
	within a ruby “column” have the measure of the widest content in that “column”.
	In the case of <i>spanning annotations</i>
	(whether actually spanning or pretending to span per 'ruby-merge'),
	the measures of the <i>ruby annotation box</i> and
	the sum of its associated <i>ruby base boxes</i> must match.

	<p>How the extra space is distributed
	when ruby content is narrower than the measure of its box
	is specified by the 'ruby-align' property.

<h4 id="inter-character-layout">
Inter-character ruby layout</h4>

	<p>Inter-character annotations have special layout.
	When 'ruby-position' indicates ''inter-character'' annotations,
	the affected <i>ruby annotation boxes</i>
	are spliced into and measured as part of the layout of the <i>base level</i>.
	The <i>ruby base container</i> must be sized to include both the <i>ruby base boxes</i>
	as well as the ''inter-character'' <i>ruby annotation boxes</i>.
	The affected <i>ruby annotation container</i> is similarly sized
	so that its content box coincides with that of the <i>ruby base container</i>.

	<p>For the purpose of laying out other levels of annotations,
	an ''inter-character'' annotation effectively becomes part of its base.
	<span class="issue">Or should it become a quasi-base between two bases?</span>
	A spanning ''inter-character'' annotation is placed after
	all the bases that it spans.

<h3 id="box-style">
Styling Ruby Boxes</h3>

	<p>In most respects, ruby boxes can be styled similar to inline boxes.
	However, the UA is not required to support
	any of the box properties (borders, margins, padding),
	any of the background properties or outline properties,
	or any other property that illustrates the bounds of the box
	on <i>ruby base container boxes</i>, <i>ruby annotation container boxes</i>,
	or <a href="#nested-pairing">ruby-internal <i>ruby container boxes</i></a>.
	The UA may implement these boxes simply as abstractions for inheritance
	and control over the layout of their contents.

<h3 id="line-breaks">
Ruby box and line breaking</h3>

	<p>When there is not enough space for an entire <i>ruby container</i> to fit on the line,
	the ruby may be broken wherever all levels simultaneously allow a break.
	Ruby most often breaks between base-annotation sets,
	but if the line-breaking rules allow it, can also break within a <i>ruby base</i>
	(and, in parallel, its associated <i>ruby annotation boxes</i>).

	<p>Whenever ruby breaks across lines, <i>ruby annotations</i> must stay
	with their respective <i>ruby bases</i>.
	The line <em>must not</em> break between a <i>ruby base</i> and its <i>annotations</i>,
	even in the case of ''inter-character'' <i>annotations</i>.

	<div class="figure">
		<img src="images/r-break-b.gif"
		     alt='Diagram showing the line breaking opportunity in a "Bopomofo" ruby'>
		<p class="caption">''inter-character'' ruby line breaking opportunity
	</div>

<h4 id="break-between">
Breaking between bases</h4>

	<p>In typical cases, <i>ruby base boxes</i> and <i>ruby annotation boxes</i>
	are styled to forbid internal line wrapping and do not contain forced breaks.
	(See <a href="#default-stylesheet">Appendix A</a>.)
	In such cases the <i>ruby container</i> can only break between adjacent <i>ruby bases</i>,
	and only if no <i>ruby annotations</i> span those <i>ruby bases</i>.

	<div class="figure">
		<p><img src="images/r-break-a.gif"
		     alt="Diagram showing the line breaking opportunity in a complex ruby">
		<p class="caption">Ruby line breaking opportunity
	</div>

	<p>Whether ruby can break between two adjacent <i>ruby bases</i>
	is controlled by normal line-breaking rules for the base text,
	exactly as if the <i>ruby bases</i> were adjacent inline boxes.
	(The annotations are ignored when determining soft wrap opportunities for the <i>base level</i>.)

	<div class="example">
		<p>For example, if two adjacent ruby bases are “蝴” and “蝶”,
		the line may break between them,
		because lines are normally allowed to break between two Han characters.
		However, if 'word-break' is ''keep-all'', that line break is forbidden.
		<pre>&lt;ruby>蝴&lt;rt>hú&lt;/rt>蝶&lt;rt>dié&lt;/rt></pre>
	</div>

	<p>Inter-base white space is significant for evaluating line break opportunities between <i>ruby bases</i>.
	As with white space between inlines, it collapses when the line breaks there.
	Similarly, annotation white space is also trimmed at a line break.

	<div class="example">
		<p>For example, given the following markup:
		<pre>&lt;ruby>&lt;rb>one&lt;/rb> &lt;rb>two&lt;/rb> &lt;rt>1&lt;/rt> &lt;rt>2&lt;/rt>&lt;/ruby></pre>
		<p>Due to the space, the line may break between “one” and “two“.
		If the line breaks there, that space&mdash;and the space between “1” and “2”&mdash;disappears,
		in accordance with standard CSS white space processing rules. [[CSS3TEXT]]
	</div>

<h4 id="break-within">
Breaking within bases</h4>

	<p>For longer base texts, it is sometimes appropriate to allow breaking within a base-annotation pair.
	For example, if an English sentence is annotated with its Japanese translation,
	allowing the text to wrap allows for reasonable line breaking behavior in the paragraph.

	<p class="issue">
	Insert scanned example so people don't think this is just the ramblings of an insane spec-writer.

	<p>Line-breaking within a <i>ruby base</i> is only allowed if the 'white-space' property
	of the <i>ruby base</i> and all its parallel <i>annotations</i> allow it,
	and there exists a <i>soft wrap opportunity</i> <em>within</em> (i.e. not at the start or end)
	the content of each base/annotation box.
	Since there is no structural correspondence between fragments of content
	within <i>ruby bases</i> and <i>annotations</i>,
	the UA may break at any set of opportunities;
	but it is recommended that the UA attempt to proportionally balance
	the amount of content inside each fragment.

	<p>There are no line breaking opportunities within ''inter-character'' <i>annotations</i>.

	<p>Ruby alignment takes place within each fragment, after line-breaking.

<h3 id="ruby-bidi">
Bidi Reordering</h3>

	<p class="issue">
	Constraints:
	Text within a ruby base must remain contiguous,
	and bases belonging to a single annotation must remain contiguous.
	Still figuring out exactly to enforce these limitations in a sensible manner.

<!--

	<p>The Unicode bidirectional algorithm orders logically-stored text for visual presentation
	when characters from scripts of opposing directionalities are mixed
	within a single paragraph.
	(See [[CSS3-WRITING-MODES]] for a more in-depth discussion of bidirectional text in CSS.)
	Bidi reordering of ruby-annotation pairs is controlled by the ordering of the base text, as follows:

<h4 id="bidi-B">Proposal A</h4>

	<p class="note">This proposal is simpler, but will require more frequent tagging of mixed-directionality content.

	<p>To avoid the interference of the <i>annotations</i> in the ordering of base text,
	all annotations are ignored for the purpose of resolving neutral characters in the base level.

	<p>To preserve the correspondance of <i>ruby annotations</i>
	to their respective <i>ruby bases</i>,
	a few restrictions are imposed:
	<ul>
		<li>The contents of a <i>ruby base</i> or <i>ruby annotation</i> must remain contiguous.
			To this end, the contents of each <i>ruby base</i> and <i>ruby annotation</i> are treated as if within a <i>bidi isolation</i>.
		<li>For the purpose of ordering <i>ruby bases</i> within a <i>ruby container</i>,
			each <i>ruby base</i> is treated as a strong character of its specified 'direction'.
			(<i>Ruby annotations</i> are positioned with respect to their <i>bases</i>,
			so are affected by reordering, but do not themselves participate in reordering.)
		<li>All <i>bases</i> spanned by a single <i>annotation</i> must remain contiguous.
			To this end, the embedding level of all bases spanned by a spanning <i>annotation</i>
			is increased by two prior to reordering.
	</ul>

<h4 id="bidi-B">Proposal B</h4>

	<p class="note">This preserves some aspects of implicit bidi.
	For example, annotating each half of "first-second" would not cause the word to reverse itself to "second-first" within an opposite-order paragraph.

	<p>To avoid the interference of the <i>annotations</i> in the ordering of the base text,
	all annotations are ignored for the purpose of resolving neutral characters in the base text;
	and the base text is resolved exactly as if each <i>ruby base</i> were
	just a normal <i>inline box</i> embedded in the <i>ruby container</i>'s inline formatting context.

	<p>Furthermore, to preserve the correspondance of <i>ruby annotations</i>
	to their respective <i>ruby bases</i>,
	a few restrictions are imposed:
	<ul>
		<li>The contents of a <i>ruby base</i> must remain contiguous.
		<li>All <i>bases</i> spanned by a single <i>annotation</i> must remain contiguous.
	</ul>
	<p>To this end, the text of <i>ruby base</i> boxes that are not <i>bidi-isolated</i>
	must have its embedding level increased by two;
	and if an <i>annotation</i> spans more than one <i>ruby base</i>,
	the spanned text's embedding level
	(or the embedding level assigned to the isolation, if the <i>ruby base</i> is <i>bidi-isolated</i>)
	is further increased by two.

-->

<h3 id="line-height">
Ruby box and line stacking</h3>

	<p>The 'line-height' property controls spacing between lines in CSS.
	When inline content on line is shorter than the 'line-height',
	half-leading is added on either side of the content,
	as specified in <a href="http://www.w3.org/TR/CSS21/visudet.html#line-height">CSS2.1&sect;10.8</a>. [[!CSS21]]

	<p>In order to ensure consistent spacing of lines,
	documents with ruby typically ensure that the 'line-height' is large enough
	to accommodate ruby between lines of text.
	Therefore, ordinarily, <i>ruby annotation containers</i> and <i>ruby annotation boxes</i>
	do not contribute to the measured height of a line's inline contents;
	any alignment (see 'vertical-align') and line-height calculations
	are performed using only the <i>ruby base container</i>,
	exactly as if it were a normal inline.

	<p>However, if the 'line-height' specified on the <i>ruby container</i>
	is less than the distance between
	the top of the top <i>ruby annotation container</i>
	and the bottom of the bottom <i>ruby annotation container</i>,
	then additional leading is added
	on the appropriate side of the <i>ruby base container</i>
	such that if a block consisted of three lines
	each containing ruby identical to this,
	none of the <i>ruby containers</i> would overlap.

	<p class="note">Note that this does not ensure that the <i>ruby annotations</i> remain within the line box.
	It merely ensures that <em>if all lines had equal spacing</em>
	and equivalent amounts and positioning of <i>ruby annotations</i>,
	there would be enough room to avoid overlap.

	<p>Authors should ensure appropriate 'line-height' and 'padding' to accommodate ruby,
	and be particularly careful at the beginning or end of a block
	and when a line contains inline-level content
	(such as images, inline blocks, or elements shifted with 'vertical-align')
	taller than the paragraph's default font size.

	<div class="figure">
		<p><img src="images/rlh-a.gif"
		        alt="The content of each line sits in the middle of its line height;
		             the additional space on each side is called half-leading.
		             Ruby fits between lines if it is smaller than twice the half-leading,
		             but this means that it occupies space belonging to the half-leading of the previous line.">
		<p class="caption">Ruby annotations will often overflow the line;
		authors should ensure content over/under a ruby-annotated line
		is adequately spaced to leave room for the ruby.
	</div>

	<p class="note">More control over how ruby affects alignment and line layout
	will be part of the CSS Line Layout Module Level 3.
	Note, it is currently in the process of being rewritten;
	the current drafts should not be relied upon.

<h2 id="ruby-props">
Ruby Properties</h2>

	<p>The following properties are introduced to control ruby positioning and alignment.

<h3 id="rubypos">
Ruby positioning: the 'ruby-position' property</h3>

	<table class="propdef">
		<tr>
			<th>Name:
			<td><dfn>ruby-position</dfn>
		<tr>
			<th><a href="#values">Value</a>:
			<td>[ over | under | inter-character ] && [ right | left ]
		<tr>
			<th>Initial:
			<td>over right
		<tr>
			<th>Applies to:
			<td>ruby annotation containers
		<tr>
			<th>Inherited:
			<td>yes
		<tr>
			<th>Percentages:
			<td>N/A
		<tr>
			<th>Media:
			<td>visual
		<tr>
			<th>Computed value:
			<td>specified value
		<tr>
			<th>Animatable:
			<td>no
		<tr>
			<th>Canonical order:
			<td><abbr title="follows order of property value definition">per grammar</abbr>
	</table>

	<p>This property controls position of the ruby text with respect to its base.
	Values have the following meanings:

	<p class="issue"><span class="issuehead">Issue-107:&nbsp;</span> Roland Steiner has requested the addition of an auto value as default. See <a href="http://www.w3.org/Search/Mail/Public/advanced_search?keywords=&amp;hdr-1-name=subject&amp;hdr-1-query=ruby-position%3A+undesirable+default+value+%27before%27+for+complex+ruby&amp;hdr-2-name=from&amp;hdr-2-query=&amp;hdr-3-name=message-id&amp;hdr-3-query=&amp;period_month=&amp;period_year=&amp;index-grp=Public__FULL&amp;index-type=t&amp;type-index=www-style&amp;resultsperpage=20&amp;sortby=date">this thread</a> and <a href="http://www.w3.org/Search/Mail/Public/advanced_search?keywords=&amp;hdr-1-name=subject&amp;hdr-1-query=Styling+of+complex+Ruby&amp;hdr-2-name=from&amp;hdr-2-query=&amp;hdr-3-name=message-id&amp;hdr-3-query=&amp;period_month=&amp;period_year=&amp;index-grp=Public__FULL&amp;index-type=t&amp;type-index=public-i18n-core&amp;resultsperpage=20&amp;sortby=date">this one</a>.</p>
	<dl dfn-for=ruby-position>
		<dt><dfn>''over''</dfn>
		<dd>The ruby text appears <i>over</i> the base in horizontal text.

			<div class="figure">
				<p><img src="images/shinkansen-top.gif"
				        alt="Diagram of ruby glyph layout in horizontal mode with ruby text appearing above the base">
				<p class="caption">Ruby over Japanese base text in horizontal layout
			</div>
		</dd>

		<dt><dfn>''right''</dfn>
		<dd>The ruby text appears on the right side of the base in vertical text.
			<div class="figure">
				<p><img src="images/shinkansen-right.gif" width="33"
				        alt="Diagram of ruby glyph layout in vertical mode with ruby text appearing vertically on the right of the base">
				<p class="caption">Ruby to the right of Japanese base text in vertical layout
			</div>
		</dd>

		<dt><dfn>''under''</dfn>
		<dd>The ruby text appears under the base in horizontal text.
			This is a relatively rare setting used in ideographic East Asian writing systems,
			most easily found in educational text.

			<div class="figure">
				<p><img src="images/shinkansen-bottom.gif"
				        alt="Diagram of ruby glyph layout in horizontal mode with ruby text appearing below the base">
				<p class="caption">Ruby under Japanese base text in horizontal layout
			</div>
		</dd>

		<dt><dfn>''left''</dfn>
		<dd>The ruby text appears on the left side of the base in vertical text.

			<div class="figure">
				<p><img src="images/shinkansen-left.gif"
				        alt="Diagram of ruby glyph layout in vertical mode with ruby text appearing vertically on the left of the base">
				<p class="caption">Ruby to the left of Japanese base text in vertical layout
			</div>
		</dd>

		<dt><dfn>''inter-character''</dfn></dt>
		<dd>
			<p>The ruby text appears on the right of the base in horizontal text.
			This value forces the 'writing-mode' of the <i>ruby annotation</i> to be vertical.
			
			<p>This value is provided for the special case of traditional Chinese
			as used especially in Taiwan:
			ruby (made of <a href="#g-bopomofo">bopomofo</a> glyphs) in that context
			appears vertically along the right side of the base glyph,
			even when the layout of the base characters is horizontal:

				<div class="figure">
					<p><img src="images/bopomofo.gif"
					        alt="Example of Taiwanese-style ruby">
					<p class="caption">“Bopomofo” ruby in traditional Chinese
					(ruby text shown in blue for clarity) in horizontal layout
				</div>
			<p class="note">
				Note that the user agent is responsible for ensuring the correct relative alignment and positioning of the glyphs,
				including those corresponding to the tone marks, when displaying.
				Tone marks are spacing characters that occur (in memory) at the end of the ruby text for each base character.
				They are usually displayed in a separate column to the right of the bopomofo characters,
				and the height of the tone mark depends on the number of characters in the syllable.
				One tone mark, however, is placed above the bopomofo, not to the right of it.
			<!-- See Taiwanese requirements doc for EPUB at http://epub-revision.googlecode.com/files/EGLS_TW_eng.ppt -->
		</dd>
	</dl>

	<p>If multiple <i>ruby annotation containers</i> have the same 'ruby-position',
	they stack along the block axis,
	with lower levels of annotation closer to the base text.

<h3 id="collapsed-ruby">
Collapsed Ruby Annotations: the 'ruby-merge' property</h3>

	<table class="propdef">
		<tr>
			<th>Name:
			<td><dfn>ruby-merge</dfn>
		<tr>
			<th><a href="#values">Value</a>:
			<td>separate | collapse | auto
		<tr>
			<th>Initial:
			<td>separate
		<tr>
			<th>Applies to:
			<td>ruby annotation containers
		<tr>
			<th>Inherited:
			<td>yes
		<tr>
			<th>Percentages:
			<td>N/A
		<tr>
			<th>Media:
			<td>visual
		<tr>
			<th>Computed value:
			<td>specified value
		<tr>
			<th>Animatable:
			<td>no
		<tr>
			<th>Canonical order:
			<td><abbr title="follows order of property value definition">per grammar</abbr>
	</table>

	<p>
		This property controls how ruby annotation boxes should be rendered
		when there are more than one in a ruby container box:
		whether each pair should be kept separate,
		the annotations should be <dfn title="collapsed annotation">collapsed</dfn> and rendered as a group,
		or the separation should be determined based on the space available.

	<p>Possible values:</p>
	<dl dfn-for=ruby-merge>
		<dt><dfn title="ruby-merge:separate">''separate''</dfn>
		<dd>
			<p>
				Each ruby annotation box is rendered in the same column(s) as its corresponding base box(es).
				This style is called “mono ruby” in [[JLREQ]].

			<div class="example">
				<p>For example, the following two markups render the same:
				<pre>&lt;ruby&gt;無&lt;rt&gt;む&lt;/ruby&gt;&lt;ruby&gt;常&lt;rt&gt;じょう&lt;/ruby&gt;</pre>
				<p>and:
				<pre>&lt;ruby style="ruby-merge:separate"&gt;&lt;rb&gt;無&lt;rb&gt;常&lt;rt&gt;む&lt;rt&gt;じょう&lt;/ruby&gt;</pre>
			</div>
		</dd>

		<dt><dfn title="ruby-merge:collapse">''collapse''</dfn>
		<dd>
			<p>
				All <i>ruby annotation boxes</i> within the same <i>ruby segment</i> on the same line are concatenated,
				and laid out as if their contents belonged to a single <i>ruby annotation box</i>
				spanning all their associated <i>ruby base boxes</i>.
				This style renders similar to “group ruby” in [[JLREQ]],
				except that <i>ruby annotations</i> are kept together with their respective <i>ruby bases</i> when breaking lines.
			</p>

			<div class="example">
				<p>The following two markups render the same both characters fit on one line:
				<pre>&lt;ruby&gt;無常&lt;rt&gt;むじょう&lt;/ruby&gt;</pre>
				<p>and:
				<pre>&lt;ruby style="ruby-merge:collapse"&gt;&lt;rb&gt;無&lt;rb&gt;常&lt;rt&gt;む&lt;rt&gt;じょう&lt;/ruby&gt;</pre>
				<p>However, the second one renders the same as ''ruby-position: separate''
				when the two bases are split across lines.
			</div>
		</dd>

		<dt><dfn title="ruby-merge:auto">''auto''</dfn></dt>
		<dd>
			<p>
				The user agent may use any algorithm to determine how each ruby annotation box
				is rendered to its corresponding base box,
				with the intention that if all annotations fit over their respective bases,
				the result is identical to “mono ruby”,
				but if some annotations are wider than their bases
				the space is shared in some way
				to avoid imposing space between bases.
			<div class="example">
			<p>
				One possible algorithm is described as “jukugo ruby” in [[JLREQ]].
			<p>
				Another, more simplified algorithm of “jukugo ruby” is
				to render as ''separate'' if all ruby annotation boxes fit
				within the advances of their corresponding base boxes,
				and render as ''collapse'' otherwise.
			</p>
			</div>
		</dd>
	</dl>

<h3 id="ruby-align-property"><span id="rubyalign"></span>
Ruby Text Distribution: the 'ruby-align' property</h3>

	<table class="propdef">
		<tr>
			<th>Name:
			<td><dfn>ruby-align</dfn>
		<tr>
			<th><a href="#values">Value</a>:
			<td>start | center | space-between | space-around
		<tr>
			<th>Initial:
			<td>space-around
		<tr>
			<th>Applies to:
			<td>ruby bases, ruby annotations, ruby base containers, ruby annotation containers
		<tr>
			<th>Inherited:
			<td>yes
		<tr>
			<th>Percentages:
			<td>N/A
		<tr>
			<th>Media:
			<td>visual
		<tr>
			<th>Computed value:
			<td>specified value (except for initial and inherit)
	</table>

	<p>This property specifies how text is distributed within the various ruby boxes
	when their contents do not exactly fill their respective boxes.
	Note that space distributed by 'ruby-align' is unrelated to, and independent of,
	any space distributed due to justification.

	<p>Values have the following meanings:
	<dl dfn-for=ruby-align>
		<dt><dfn title="ruby-align:start">''start''</dfn></dt>
		<dd>The ruby content is aligned with the start edge of its box.
			<div class="figure">
				<p><img
					alt="Diagram of glyph layout in left aligned ruby when ruby text is shorter than base"
					width="145" height="91" src="images/ra-l.gif" /><img
					width="145" height="91"
					alt="Diagram of glyph layout in left aligned ruby when ruby text is longer than base"
					src="images/ra-l-rb.gif" />
				<p class="caption">''start'' ruby distribution
			</div>
		</dd>

		<dt><dfn title="ruby-align:center">''center''</dfn></dt>
		<dd>The ruby content is centered within its box.
			<div class="figure">
				<p><img width="145" height="91"
					alt="Diagram of glyph layout in center aligned ruby when ruby text is shorter than base"
					src="images/ra-c.gif" /><img width="145" height="91"
					alt="Diagram of glyph layout in center aligned ruby when ruby text is longer than base"
					src="images/ra-c-rb.gif" />
				<p class="caption">''center'' ruby distribution
			</div>
		</dd>

		<dt><dfn title="ruby-align:space-between">''space-between''</dfn></dt>
		<dd>
			<p>The ruby content expands as defined for normal text justification
				(as defined by 'text-justify'),
				except that if there are no <i>justification opportunities</i>
				the content is centered.
			<div class="figure">
				<p><img width="145" height="91"
				alt="Diagram of glyph layout in distribute-letter aligned ruby when ruby text is shorter than base"
				src="images/ra-dl.gif" /><img width="145" height="91"
				alt="Diagram of glyph layout in distribute-letter aligned ruby when ruby text is longer than base"
				src="images/ra-dl-rb.gif" />
				<p class="caption">''space-between'' ruby distribution
			</div>
		</dd>

		<dt><dfn title="ruby-align:space-around">''space-around''</dfn></dt>
		<dd>
			<p>As for ''space-between''
			except that there exists an extra <i>justification opportunities</i>
			whose space is distributed half before and half after the ruby content.
			<div class="example">
				<p>Since a typical implementation will by default define <i>justification opportunities</i>
					between every adjacent pair of CJK <i>characters</i>
					and not between adjacent pairs of Latin <i>characters</i>,
					this should result in the behavior recommended by [[JLREQ]]:
					for wide-cell ruby content to be distributed...
				<div class="figure">
					<p><img width="145" height="91"
					alt="Diagram of glyph layout in auto aligned ruby when ruby text is shorter than base"
					src="images/ra-ds.gif" /><img width="145" height="91"
					alt="Diagram of glyph layout in auto aligned ruby when ruby text is longer than base"
					src="images/ra-ds-rb.gif" />
					<p class="caption">Wide-cell text in ''space-around'' ruby distribution is spaced apart
				</div>
				<p>... and narrow-cell glyph ruby to be centered.
				<div class="figure">
					<p><img
					alt="Diagram of glyph layout in auto aligned ruby when halfwidth ruby text is shorter than base"
					width="145" height="91"
					src="images/ra-c-h.gif" /><img
					alt="Diagram of character layout in auto aligned ruby when ruby text is longer than narrow-width base"
					width="145" height="91"
					src="images/ra-c-rb-h.gif" />
					<p class="caption">Narrow-width ruby text in ''space-around'' ruby distribution is centered
				</div>
			</div>
		</dd>
	</dl>

	<p class="issue">Add a paragraph explaining how to distribute space in situations with spanning annotations.
<!--
	<p>For a complex ruby with spanning elements, one additional consideration is
		required. If the spanning element spans multiple 'rows' (other rbc or rtc
		elements), and the ruby alignment requires space distribution among the
		'spanned' elements, a ratio must be determined among the 'columns' of spanned
		elements. This ratio is computed by taking into consideration the widest
		element within each column.</p>
-->
<h2 id="edge-effects">
Edge Effects</h2>

<h3 id="ruby-overhang">
Overhanging Ruby</h3>

	<p>
		When <i>ruby annotation box</i> is longer than its corresponding <i>ruby base box</i>,
		the <i>ruby annotation box</i> may partially overhang adjacent boxes.
	</p>
	<p>
		This level of the specification does not define
		how much the overhang may be allowed, and under what conditions.
	</p>

	<p>If the ruby text is not allowed to overhang,
	then the ruby behaves like a traditional inline box,
	i.e. only its own contents are rendered within its boundaries
	and adjacent elements do not cross the box boundary:

	<div class="figure">
		<p><img src="images/ro-n.gif"
		        alt="Diagram showing the ruby boxes interacting with adjacent text">
		<p class="caption">Simple ruby whose text is not allowed to overhang adjacent text
	</div>

	<p>However, if <i>ruby annotation</i> content is allowed to overhang adjacent elements
	and it happens to be wider than its base,
	then the adjacent content is partially rendered within the area of the <i>ruby container box</i>,
	while the <i>ruby annotation</i> may partially overlap the upper blank parts of the adjacent content:

	<div class="figure">
	<p><img src="images/ro-a.gif"
		      alt="Diagram showing the ruby boxes interacting with adjacent text">
	<p class="caption">Simple ruby whose text is allowed to overhang adjacent text
	</div>

	<p>The <i>ruby annotations</i> related to a <i>ruby base</i>
	must never overhang another <i>ruby base</i>.

	<p>The alignment of the contents of the base or the ruby text
	is not affected by overhanging behavior.
	The alignment is achieved the same way regardless of the overhang behavior setting
	and it is computed before the space available for overlap is determined.
	It is controlled by the 'ruby-align' property.

	<p class="issue">
		I suspect overhanging interacts with alignment in some cases;
		might need to look into this later.

	<p>This entire logic applies the same way in vertical ideographic layout,
	only the dimension in which it works in such a layout is vertical,
	instead of horizontal.

	<div class="example">
	<p>
		The user agent may use [[JIS4051]] recommendation of
		using one ruby text character length as the maximum overhang length.
		Detailed rules for how ruby text can overhang adjacent characters for Japanese are described by [[JLREQ]].
	</p>
	</div>

<h3 id="line-edge">
Line-edge Alignment</h3>

	<p>
		When a <i>ruby annotation box</i> that is longer than its <i>ruby base</i>
		is at the start or end edge of a line,
		the user agent <em>may</em> force the side of the <i>ruby annotation</i> that touches the edge of the line
		to align to the corresponding edge of the base.
		This type of alignment is described by [[JLREQ]].
	</p>
	<p>
		This level of the specification does not provide a mechanism to control this behavior.
	</p>
	<div class="figure">
		<p><img src="images/ra-le-l.gif"
			alt="Diagram of glyph layout in line-edge aligned ruby when ruby text is shorter than base">
			<img src="images/ra-le-r.gif"
			alt="Diagram of glyph layout in line-edge aligned ruby when ruby text is longer than base">
		<p class="caption">Line-edge alignment
	</div>

	<!--
<h3 id="rubyover">
Ruby overhanging: the 'ruby-overhang' property</h3>

  <table class="propdef">
    <tr>
      <th>Name:
      <td><dfn>ruby-overhang</dfn>
    <tr>
      <th>Value:
      <td>auto | start | end | none
    <tr>
      <th>Initial:
      <td>none
    <tr>
      <th>Applies to:
      <td>the parent of elements with display: ruby-text
    <tr>
      <th>Inherited:
      <td>yes
    <tr>
      <th>Percentages:
      <td>N/A
    <tr>
      <th>Media:
      <td>visual
    <tr>
      <th>Computed value:
      <td>specified value (except for initial and inherit)
  </table>

<p>This property determines whether, and on which side, ruby text is allowed
to partially overhang any adjacent text in addition to its own base, when the
ruby text is wider than the ruby base. Note that ruby text is never allowed to
overhang glyphs belonging to another ruby base. <span class="issue"><span class="issuehead">Issue:&nbsp;</span> This rule must be broken if we are to allow support for jukugo ruby.</span> Also the user agent is free to assume
a maximum amount by which ruby text may overhang adjacent text. The user agent may use
the [[JIS4051]] recommendation of using one ruby text character
length as the maximum overhang length. Detailed rules for how ruby text can overhang adjacent characters for Japanese are described by [[JLREQ]].</p>

<p>Possible values:</p>
<dl>
  <dt><strong>auto</strong></dt>
    <dd>The ruby text can overhang text adjacent to the base on either side.   	  [[JLREQ]] and [[JIS4051]] specify the categories of characters that
      ruby text can overhang. The user agent is free to follow those recommendations or specify its own classes of
      characters to overhang. This is the initial value.
		<div class="figure">
      <p><img class="example" width="177" height="91"
      alt="Diagram of glyph layout in overhanging ruby" src="images/ro-a.gif" /></p>
      <p><b>Figure 4.3.1</b>: Ruby overhanging adjacent text</p>
      </div>
    </dd>
  <dt><strong>start</strong></dt>
    <dd>The ruby text can only overhang the text that precedes it. That means, for
      example, that ruby cannot overhang text that is to the right of it in
      horizontal LTR layout, and it cannot overhang text that is below it in
      vertical-ideographic layout.
		<div class="figure">
      <p><img class="example" width="199" height="91"
      alt="Diagram of glyph layout when ruby overhangs the preceding glyphs only"
      src="images/ro-s.gif" /></p>
      <p><b>Figure 4.3.2</b>: Ruby overhanging preceding text only</p>
      </div>
    </dd>
  <dt><strong>end</strong></dt>
    <dd>The ruby text can only overhang the text that follows it. That means, for
      example, that ruby cannot overhang text that is to the left of it in
      horizontal LTR layout, and it cannot overhang text that is above it in
      vertical-ideographic layout.
		<div class="figure">
      <p><img class="example" width="198" height="91"
      alt="Diagram of glyph layout when ruby overhangs the following characters only"
      src="images/ro-e.gif" /></p>
      <p><b>Figure 4.3.3</b>: Ruby overhanging following text only</p>
      </div>
    </dd>
  <dt><strong>none</strong></dt>
    <dd>The ruby text cannot overhang any text adjacent to its base, only its
      own base.

      <div class="figure">
      <p><img class="example" width="220" height="91"
      alt="Diagram of glyph layout in non-overhanging ruby"
      src="images/ro-n.gif" /></p>
      <p><b>Figure 4.3.4</b>: Ruby not allowed to overhang adjacent text</p>
      </div>
    </dd>
</dl>

<h3 id="rubyspan">
Ruby annotation spanning: the 'ruby-span' property</h3>

  <table class="propdef">
    <tr>
      <th>Name:
      <td><dfn>ruby-span</dfn>
    <tr>
      <th>Value:
      <td>attr(x) |  none
    <tr>
      <th>Initial:
      <td>none
    <tr>
      <th>Applies to:
      <td>elements with display: ruby-text
    <tr>
      <th>Inherited:
      <td>no
    <tr>
      <th>Percentages:
      <td>N/A
    <tr>
      <th>Media:
      <td>visual
    <tr>
      <th>Computed value:
      <td>&lt;number&gt;
  </table>

<p>This property controls the spanning behavior of annotation elements. </p>

<p class="note"><span class="note-label">Note:</span> A XHTML user agent may also use the <samp>rbspan</samp> 
attribute to get the same effect.</p>

<p>Possible values:</p>

<dl>
  <dt><strong>attr(x)</strong></dt>
    <dd>The value of attribute 'x' as a string value. The string value is 
    evaluated as a &lt;number&gt; to determine the number of ruby base elements to be 
    spanned by the annotation element. If the &lt;number&gt; is &#39;0&#39;, it is replaced by 
    &#39;1&#39;.The &lt;number&gt; is the computed value. </dd>
  <dt>none</dt>
  <dd>No spanning. The computed value is &#39;1&#39;.</dd>
</dl>

<p>The following example shows an XML example using the 'display' property 
values associated with the 'ruby structure and the 'ruby-span' property</p>
<pre class="xml">myruby       { display: ruby; }
myrbc        { display: ruby-base-container; }
myrb         { display: ruby-base; }
myrtc.before { display: ruby-text-container; ruby-position: before}
myrtc.after  { display: ruby-text-container; ruby-position: after}
myrt         { display: ruby-text; ruby-span: attr(rbspan); }
...
&lt;myruby&gt;
  &lt;myrbc&gt;
    &lt;myrb&gt;10&lt;/myrb&gt;
    &lt;myrb&gt;31&lt;/myrb&gt;
    &lt;myrb&gt;2002&lt;/myrb&gt;
  &lt;/myrbc&gt;
  &lt;myrtc class=&quot;before&quot;&gt;
    &lt;myrt&gt;Month&lt;/myrt&gt;
    &lt;myrt&gt;Day&lt;/myrt&gt;
    &lt;myrt&gt;Year&lt;/myrt&gt;
  &lt;/myrtc&gt;
  &lt;myrtc class=&quot;after&quot;&gt;
    &lt;myrt rbspan=&quot;3&quot;&gt;Expiration Date&lt;/myrt&gt;
  &lt;/myrtc&gt;
&lt;/myruby&gt;</pre>
	-->

<h2 id="default-stylesheet" class="no-num">
Appendix A: Default Style Sheet</h2>

	<p><em>This section is informative.</em>

<h3 id="default-ua-ruby" class="no-num">
<span class="secno">A.1</span> Supporting Ruby Layout</h3>

	<p>The following represents a default UA style sheet
	for rendering HTML and XHTML ruby markup as ruby layout:

	<pre>
<!--	-->ruby { display: ruby; }
<!--	-->rb   { display: ruby-base; white-space: nowrap; }
<!--	-->rp   { display: none; }
<!--	-->rt   { display: ruby-text; white-space: nowrap; font-size: 50%; }
<!--	-->rbc  { display: ruby-base-container; }
<!--	-->rtc  { display: ruby-text-container; }
<!--	-->ruby, rb, rt, rbc, rtc { unicode-bidi: isolate; }</pre>

	<p>Additional rules for UAs supporting the relevant features of [[CSS3-TEXT-DECOR]] and [[CSS3-FONTS]]:
	<pre>rt { font-variant-east-asian: ruby; text-emphasis: none; }</pre>

	<p class="note">Authors should not use the above rules;
	a UA that supports ruby layout should provide these by default.

<h3 id="default-inline" class="no-num">
<span class="secno">A.2</span> Inlining Ruby Annotations</h3>

	<p>The following represents a sample style sheet
	for rendering HTML and XHTML ruby markup as inline annotations:

	<pre>ruby, rb, rt, rbc, rtc, rp {
<!--	-->  display: inline; white-space: inherit;
<!--	-->  font-variant-east-asian: inherit; text-emphasis: inherit; }</pre>

<h3 id="default-parens" class="no-num">
<span class="secno">A.3</span> Generating Parentheses</h3>

	<p>Unfortunately, because Selectors cannot match against text nodes,
	it's not possible with CSS to express rules that will automatically and correctly
	add parentheses to unparenthesized ruby annotations in HTML.
	(This is because HTML ruby allows implying the <i>ruby base</i> from raw text, without a corresponding element.)
	However, these rules will handle cases where either <code>&lt;rb&gt;</code>
	or <code>&lt;rtc&gt;</code> is used rigorously.

	<pre>
<!--	-->/* Parens around &lt;rtc> */
<!--	-->rtc::before { content: "("; }
<!--	-->rtc::after  { content: ")"; }

<!--	-->/* Parens before first &lt;rt> not inside &lt;rtc> */
<!--	-->rb  + rt::before,
<!--	-->rtc + rt::before { content: "("; }

<!--	-->/* Parens after &lt;rt> not inside &lt;rtc> */
<!--	-->rb ~ rt:last-child::after,
<!--	-->rt + rb::before  { content: ")"; }
<!--	-->rt + rtc::before { content: ")("; }</pre>

<h2 id="glossary">
Glossary</h2>
<dl>
  <dt id="g-bopomofo" lang="zh">Bopomofo
    <dd>37 characters and 4 tone markings used as phonetics in Chinese,
      especially standard Mandarin.</dd>
  <dt id="g-hanja" lang="ko">Hanja
    <dd>Subset of the Korean writing system that utilizes ideographic
      characters borrowed or adapted from the Chinese writing system. Also see
      <a href="#g-kanji"><span lang="ja">Kanji</span></a>.</dd>
  <dt id="g-hiragana" lang="ja">Hiragana
    <dd>Japanese syllabic script, or character of that script. Rounded and 
    cursive in appearance. Subset of the Japanese writing system, used together 
    with kanji and katakana. In recent times, mostly used to write Japanese 
    words when kanji are not available or appropriate, and word endings and 
    particles. Also see <a
      href="#g-katakana"><span lang="ja">Katakana</span></a>.</dd>
  <dt id="g-ideogram">Ideograph
    <dd>A character that is used to represent an idea, word, or word component, 
    in contrast to a character from an alphabetic or syllabic script. The most 
    well-known ideographic script is used (with some variation) in East Asia 
    (China, Japan, Korea,...).</dd>
  <dt id="g-kana" lang="ja">Kana
    <dd>Collective term for hiragana and katakana.</dd>
  <dt id="g-kanji">Kanji
    <dd>Japanese term for ideographs; ideographs used in Japanese. Subset of the 
    Japanese writing system, used together with hiragana and katakana. Also see <a
      href="#g-hanja"><span lang="ko">Hanja</span></a>.</dd>
  <dt id="g-katakana" lang="ja">Katakana
    <dd>Japanese syllabic script, or character of that script. Angular in 
    appearance. Subset of the Japanese writing system,&nbsp; used together with 
    kanji and hiragana. In recent times, mainly used to write foreign words. Also see <a
      href="#g-hiragana"><span lang="ja">Hiragana</span></a>.</dd>
</dl>

<h2 class=no-num id="acknowledgments">
Acknowledgments</h2>

	<p>This specification would not have been possible without the help from:</p>

	<p>
	David Baron,
	Robin Berjon,
	Stephen Deach,
	Martin Dürst,
	Hideki Hiura (<span lang="ja">樋浦 秀樹</span>),
	Masayasu Ishikawa (<span lang="ja">石川雅康</span>),
	Taichi Kawabata,
	Chris Pratley,
	Takao Suzuki (<span lang="ja">鈴木 孝雄</span>),
	Frank Yung-Fong Tang,
	Chris Thrasher,
	Masafumi Yabe (<span lang="ja">家辺勝文</span>),
	Steve Zilles.

	<p>Special thanks goes to the previous editors:
	Michel Suignard and Marcin Sawicki of Microsoft,
	and Richard Ishida of W3C.

<h2 class="no-num" id="changes">
Changes</h2>

	<p>The following major changes have been made since the previous Working Draft:
	<ul>
		<li>Improve anonymous box generation and white space handling rules.
	</ul>