HTML is the core language of the World Wide Web. The W3C publishes HTML5 and HTML5.1. The WHATWG publishes HTML, which is a rough superset of W3C HTML5.1. "Differences from HTML4" describes the differences of the HTML specifications from those of HTML4, and calls out cases where they differ from each other. This document may not provide accurate information, as the specifications are still actively in development. When in doubt, always check the specifications themselves. [HTML5] [HTML5NIGHTLY] [HTML]
This section describes the status of this document at the time of its publication. Other documents may supersede this document. A list of current W3C publications and the latest revision of this technical report can be found in the W3C technical reports index at http://www.w3.org/TR/.
This is the 28 May 2013 W3C Working Draft produced by the HTML Working Group, part of the HTML Activity. The Working Group intends to publish this document as a Working Group Note. The appropriate forum for comments is W3C Bugzilla. (public-html-comments@w3.org, a mailing list with a public archive, is no longer used for tracking comments.)
Publication as a Working Draft does not imply endorsement by the W3C Membership. This is a draft document and may be updated, replaced or obsoleted by other documents at any time. It is inappropriate to cite this document as other than work in progress.
This document was produced by a group operating under the 5 February 2004 W3C Patent Policy. W3C maintains a public list of any patent disclosures made in connection with the deliverables of the group; that page also includes instructions for disclosing a patent. An individual who has actual knowledge of a patent which the individual believes contains Essential Claim(s) must disclose the information in accordance with section 6 of the W3C Patent Policy.
This document covers the W3C HTML5 specification, W3C HTML5.1 specification, and the WHATWG HTML standard. For readability, these are referred to as if they were a single specification: "the HTML specification" or simply "HTML" when something applies equally to all of them; otherwise, they are called out explicitly.
HTML has been in continuous evolution since it was introduced to the Internet in the early 1990s. Some features were introduced in specifications; others were introduced in software releases. In some respects, implementations and Web developer practices have converged with each other and with specifications and standards, but in other ways, they have diverged.
HTML4 became a W3C Recommendation in 1997. While it continues to serve as a rough guide to many of the core features of HTML, it does not provide enough information to build implementations that interoperate with each other and, more importantly, with Web content. The same goes for XHTML1, which defines an XML serialization for HTML4, and DOM Level 2 HTML, which defines JavaScript APIs for both HTML and XHTML. HTML replaces these documents. [DOM2HTML] [HTML4] [XHTML1]
The HTML specification reflects an effort, started in 2004, to study contemporary HTML implementations and Web content. The specification:
Defines a single language called HTML which can be written in HTML syntax and in XML syntax.
Defines detailed processing models to foster interoperable implementations.
Improves markup for documents.
Introduces markup and APIs for emerging idioms, such as Web applications.
The contents of HTML, as well as the contents of this document which depend on HTML, are still being discussed on the HTML Working Group and WHATWG mailing lists.
Open issues for WHATWG HTML, see "Issues:" at the top of the specification. The specification also has annotation boxes in the margin which can link to bugs.
Open issues for W3C HTML5 and W3C HTML5.1, see the "Status of This Document" section.
HTML is defined in a way that is backward compatible with the way user agents handle content. To keep the language relatively simple for Web developers, several older elements and attributes are not included, as outlined in the other sections of this document, such as presentational elements that are better handled using CSS.
User agents, however, will always have to support these older
elements and attributes. This is why the HTML specification clearly
separates requirements for Web developers (referred to as "authors" in the
specification) and user agents; for instance, this
means that Web developers cannot use the isindex or the
plaintext element, but user agents are required to support
them in a way that is compatible with how these elements need to behave
for compatibility with Web content.
Since HTML has separate conformance requirements for Web developers and user agents there is no longer a need for marking features "deprecated".
The HTML4 specification reached Recommendation status before it was completely implemented in user agents. HTML4 still is not completely implemented, because it contains various bugs that have been fixed in the current HTML specification, which user agents are much closer to implementing than HTML4.
The WHATWG HTML standard is under continual development where bugs are fixed and new features are introduced over time. Features can also be removed from the specification if they lack implementor interest, are not being used by Web developers, or for other reasons. The WHATWG does not publish frozen snapshots.
The W3C HTML5 specification is trying to reach Recommendation status. New features are generally not added unless they are implemented by at least two browsers and have tests demonstrating interoperability. Minor bug fixes can be applied. This means that W3C HTML5 may contain bugs that have been fixed in WHATWG HTML or W3C HTML5.1, or both.
The W3C HTML5.1 specification is similar to WHATWG HTML: it is under continual development where bugs are fixed and new features are introduced over time, and features can be removed from the specification for the same reasons as with WHATWG HTML. It cherry-picks changes from the WHATWG HTML standard and also gets direct changes. It is expected to eventually go through the same procedure as W3C HTML5, and then a new version will be minted.
HTML defines a syntax, referred to as "the HTML syntax", that is
mostly compatible with HTML4 and XHTML1 documents published on the
Web, but is not compatible with the more esoteric SGML features of
HTML4, such as
processing instructions
and
shorthand markup
as these are not supported by most user agents. Documents using the HTML
syntax are served with the text/html media
type.
HTML also defines detailed parsing rules (including "error
handling") for this syntax which are largely compatible with HTML4-era
implementations. User agents have to use these rules for resources that
have the text/html media type. Here is an example document
that conforms to the HTML syntax:
<!doctype html>
<html>
<head>
<meta charset="UTF-8">
<title>Example document</title>
</head>
<body>
<p>Example paragraph</p>
</body>
</html>
The other syntax that can be used for HTML is XML. This syntax
is compatible with XHTML1 documents and implementations. Documents
using this syntax need to be served with an XML media type (such as
application/xhtml+xml or application/xml) and elements
need to be put in the http://www.w3.org/1999/xhtml
namespace following the rules set forth by the XML specifications.
[XML] [XMLNS]
Below is an example document that conforms to the XML syntax of HTML.
<?xml version="1.0" encoding="UTF-8"?>
<html xmlns="http://www.w3.org/1999/xhtml">
<head>
<title>Example document</title>
</head>
<body>
<p>Example paragraph</p>
</body>
</html>
For the HTML syntax, Web developers are required to declare the character encoding. There are three ways to do that:
At the transport level; for instance, by using the HTTP Content-Type
header.
Using a Unicode Byte Order Mark (BOM) character at the start of the file. This character provides a signature for the encoding used.
Using a meta element with a charset
attribute that specifies the encoding within the first 1024 bytes of
the document; for instance, <meta charset="UTF-8">
could be used to specify the UTF-8 encoding. This replaces the need
for
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
although that syntax is still allowed.
For the XML syntax, Web developers have to use the rules as set forth in the XML specification to set the character encoding.
The HTML syntax requires a doctype to be specified to ensure that the browser renders the page in standards mode. The doctype has no other purpose. [DOCTYPE]
The doctype declaration for the HTML syntax is <!DOCTYPE html> and is
case-insensitive. Doctypes from earlier versions of
HTML were longer because the HTML language was SGML-based and therefore
required a reference to a DTD. This is no longer the case and
the doctype is only needed to enable standards mode for documents
written using the HTML syntax. Browsers already do this for
<!DOCTYPE html>.
To support legacy markup generators that cannot generate the preferred
short doctype, the doctype <!DOCTYPE html SYSTEM
"about:legacy-compat"> is allowed in the HTML syntax.
The strict doctypes for HTML 4.0, HTML 4.01, XHTML 1.0 as well as XHTML 1.1 are also allowed (but are discouraged) in the HTML syntax.
In the XML syntax, any doctype declaration may be used, or it may be omitted altogether. Documents with an XML media type are always handled in standards mode.
The HTML syntax allows for MathML and SVG elements to
be used inside a document. An math or svg start
tag causes the HTML parser to switch to a special insertion mode which puts
elements and attributes in the appropriate namespaces, does case fixups for
elements and attributes that have mixed case, and supports the empty-element
syntax as in XML. The syntax is still case-insensitive and attributes allow
the same syntax as for HTML elements. Namespace declarations may be omitted.
CDATA sections are supported in this insertion mode.
Some MathML and SVG elements cause the parser to switch back to "HTML
mode", e.g. mtext and foreignObject, so you can
use HTML elements or a new math or svg element.
For instance, a very simple document using some of the minimal syntax features could look like:
<!doctype html> <title>SVG in text/html</title> <p> A green circle: <svg> <circle r="50" cx="50" cy="50" fill="green"/> </svg> </p>
There are a few other changes in the HTML syntax worthy of mentioning:
The ⟨ and ⟩ named character references now expand to U+27E8 and U+27E9 (mathematical left/right angle
bracket) instead of U+2329 and U+232A (left/right-pointing angle bracket), respectively.
Many new named character references have been added, including all named character references from MathML.
Void elements (known as "EMPTY" in HTML4) are allowed to have a trailing slash.
The ampersand (&) may be left unescaped in more
cases compared to HTML4.
Attributes have to be separated by at least one whitespace character.
Attributes with an empty value may be written as just the attribute name omitting the equals sign and the value, even if the attribute is not a boolean attribute. (It is commonly believed that HTML4 allowed the value to be omitted for boolean attributes. Instead, HTML4 allowed using only the attribute value and omitting the attribute name, for enumerated attributes, but this was not supported in browsers.)
Attributes omitting quotes for the value are allowed to use a larger set of characters compared to HTML4.
The HTML parser does not do any normalization of whitespace in attribute values; for instance, leading and trailing whitespace in the
id attribute is not ignored (and thus now invalid), and newline characters can be used in the
value attribute of the input element without using character
references.
The optgroup end tag is now optional.
The colgroup start tag is now optional and is inferred by
the HTML parser.
This section is split up in several subsections to more clearly illustrate the various differences from HTML4.
The following elements have been introduced for better structure:
section
represents a generic document or application section. It can be
used together
with the h1, h2, h3,
h4, h5, and h6 elements to
indicate the document structure.
article
represents an independent piece of content of a document, such as a
blog entry or newspaper article.
main
can be used as a container for the dominant contents of another element, such as
the main content of the page. In W3C HTML5 and W3C HTML 5.1, only one such element is allowed in a document.
aside
represents a piece of content that is only slightly related to the
rest of the page.
In WHATWG HTML, hgroup
represents the header of a section.
header
represents a group of introductory or navigational aids.
footer
represents a footer for a section and can contain information about
the author, copyright information, etc.
nav
represents a section of the document intended for navigation.
figure
represents a piece of self-contained flow content, typically
referenced as a single unit from the main flow of the document.
<figure> <video src="example.webm" controls></video> <figcaption>Example</figcaption> </figure>
figcaption can be used as caption (it is optional).
Then there are several other new elements:
video and
audio for multimedia content. Both
provide an API so application Web developers can script their own user
interface, but there is also a way to trigger a user interface
provided by the user agent. source
elements are used together with these elements if there are multiple
streams available of different types.
embed is used for plugin
content.
mark represents a run of
text in one document marked or highlighted for reference purposes, due
to its relevance in another context.
progress represents a
completion of a task, such as downloading or when performing a series
of expensive operations.
meter represents a
measurement, such as disk usage.
time represents a date
and/or time.
WHATWG HTML and W3C HTML5.1 has data, which
allows content to be annotated with a machine-readable value.
dialog for showing a dialog.
bdi represents a span of
text that is to be isolated from its surroundings for the purposes of
bidirectional text formatting.
wbr represents a line break
opportunity.
canvas is used for rendering
dynamic bitmap graphics on the fly, such as graphs or games.
menuitem
represents a command the user can invoke from a popup menu.
details
represents additional information or controls which the user can
obtain on demand. The
summary
element provides its summary, legend, or caption.
datalist
together with the a new list attribute for
input can be used to make comboboxes:
<input list="browsers"> <datalist id="browsers"> <option value="Safari"> <option value="Internet Explorer"> <option value="Opera"> <option value="Firefox"> </datalist>
keygen
represents control for key pair generation.
output
represents some type of output, such as from a calculation done
through scripting.
The input element's type attribute now has the
following new values:
The idea of these new types is that the user agent can provide the user interface, such as a calendar date picker or integration with the user's address book, and submit a defined format to the server. It gives the user a better experience as his input is checked before sending it to the server meaning there is less time to wait for feedback.
Several attributes have been introduced to various elements that were already part of HTML4:
The a and
area elements have the new
download attribute
in WHATWG HTML and W3C HTML5.1. WHATWG HTML also has the
ping attribute.
The area element, for consistency with the
a and link elements, now also has the
hreflang, type and rel attributes.
The base element can now have a target
attribute as well, mainly for consistency with the
a element. (This is already widely supported.)
The meta element has a charset
attribute now as this was already widely supported and provides a nice
way to specify the
character encoding for the document.
In WHATWG HTML and W3C HTML5.1, the table
element now has a sortable attribute and the
th element has a
sorted attribute, which provide a means to sort
table columns.
A new autofocus attribute can be specified on the
input (except when the type attribute is
hidden), select, textarea and
button elements. It provides a declarative way to focus a
form control during page load. Using this feature should enhance the
user experience compared to focusing the element with script as the user can turn it off if the user does not like
it, for instance.
A new placeholder attribute can be specified on
the input and textarea elements. It
represents a hint intended to aid the user with data entry.
<input type=email placeholder="a@b.com">
The new form attribute for input,
output, select, textarea,
button, label, object and fieldset elements allows for
controls to be associated with a form. These elements can now be
placed anywhere on a page, not just as descendants of the
form element, and still be associated with a form.
<table>
<tr>
<th>Key
<th>Value
<th>Action
<tr>
<td><form id=1><input name=1-key></form>
<td><input form=1 name=1-value>
<td><button form=1 name=1-action value=save>✓</button>
<button form=1 name=1-action value=delete>✗</button>
...
</table>
The new required attribute applies to
input (except when the type attribute is
hidden, image or some button type such as
submit), select and textarea. It indicates that the user
has to fill in a value in order to submit the form. For select, the first option element has to be a placeholder with an empty value.
<label>Color: <select name=color required> <option value="">Choose one <option>Red <option>Green <option>Blue </select></label>
The fieldset element now allows the
disabled attribute which disables all descendant controls (excluding those that are descendants of the legend element)
when specified, and the name attribute which can be used for script access.
The input element has several new attributes to
specify constraints: autocomplete, min,
max, multiple, pattern and
step. As mentioned before it also has a new
list attribute which can be used together with the
datalist element. It also now has the width and height attributes to specify the dimensions of the image when using type=image.
The input and textarea elements have
a new attribute named dirname that causes the
directionality of the control as set by the user to be submitted as
well.
The textarea element also has two new attributes,
maxlength and wrap which control max input
length and submitted line wrapping behavior, respectively.
The form element has a novalidate
attribute that can be used to disable form validation submission (i.e.
the form can always be submitted).
The input and button elements have
formaction, formenctype,
formmethod, formnovalidate, and
formtarget as new attributes. If present, they override
the action, enctype, method,
novalidate, and target attributes on the
form element.
In WHATWG HTML and W3C HTML5.1, the input and textarea have an inputmode attribute.
The menu element has two new attributes:
type and label. They
allow the element to transform into a menu as found in typical user
interfaces as well as providing for context menus in conjunction with the
global contextmenu attribute.
In WHATWG HTML and W3C HTML5.1, the button
element has a new menu
attribute, used together with popup menus.
The style element has a new scoped
attribute which can be used to enable scoped style sheets. Style rules
within such a style element only apply to the local tree.
The script element has a new attribute called
async that influences script loading and execution.
The html element has a new attribute called
manifest that points to an application cache manifest
used in conjunction with the API for offline Web applications.
The link element has a new attribute called
sizes. It can be used in conjunction with the
icon relationship (set through the rel
attribute; can be used for e.g. favicons) to indicate the size of the
referenced icon, thus allowing for icons of distinct dimensions.
The ol element has a new attribute called
reversed. When present, it indicates that the list order
is descending.
The iframe element has three new attributes called
sandbox, seamless, and srcdoc
which allow for sandboxing content, e.g. blog comments.
The object element has a new attribute called typemustmatch which allows safer embedding of external resources.
The img element has a new attribute called
crossorigin to use CORS in the fetch and if it is successful, allows the image data to be
read with the canvas API. In WHATWG HTML and W3C HTML5.1, the script element has a
crossorigin attribute to allow script
errors to be reported to onerror
with information about the error. WHATWG HTML and W3C HTML5.1 also has the crossorigin
attribute on the link element.
In WHATWG HTML, the img element has a new attribute called srcset to support multiple images for different resolutions and different images for different viewport sizes.
Several attributes from HTML4 now apply to all elements. These
are called global attributes: accesskey, class, dir,
id, lang, style,
tabindex and title. Additionally, XHTML 1.0
only allowed xml:space on some elements, which is now allowed
on all elements in XHTML documents.
There are also several new global attributes:
The contenteditable attribute indicates that
the element is an editable area. The user can change the contents of the
element and manipulate the markup.
The contextmenu attribute can be used to point
to a context menu provided by the Web developer.
The data-* collection of Web developer-defined
attributes. Web developers can define any attribute they want as long as they
prefix it with data- to avoid clashes with future versions of
HTML. These are intended to be used to store custom data to be consumed by
the Web page or application itself. They are not intended for
data to be consumed by other parties (e.g. user agents).
The draggable and dropzone attributes can be used together with
the new drag & drop API.
The hidden
attribute indicates that an element is not yet, or is no longer, relevant.
WHATWG HTML and W3C HTML5.1 has the inert attribute, intended to make dialog elements modal.
The role and
aria-*
collection attributes which can be used to instruct assistive
technology. These attributes have slightly different requirements in WHATWG HTML and W3C HTML5/W3C HTML5.1.
The spellcheck attribute allows for hinting
whether content can be checked for spelling or not.
The translate attribute gives a hint to
translators whether the content should be translated.
HTML also makes all event handler attributes from HTML4, which take the
form onevent, global attributes and adds
several new event handler attributes for new events it defines; for
instance, the onplay
event handler attribute for the play event which is used by the API for the
media elements (video and audio).
These elements have slightly modified meanings in HTML to better reflect how they are used on the Web or to make them more useful:
The address element is now
scoped by the nearest ancestor article
or body element.
The b element now represents a
span of text to which attention is being drawn for utilitarian purposes
without conveying any extra importance and with no implication of an
alternate voice or mood, such as key words in a document abstract, product
names in a review, actionable words in interactive text-driven software,
or an article lede.
The cite element now solely
represents the title of a work (e.g. a book, a paper, an essay, a poem, a
score, a song, a script, a film, a TV show, a game, a sculpture, a
painting, a theatre production, a play, an opera, a musical, an
exhibition, a legal case report, etc). Specifically the example in HTML4
where it is used to mark up the name of a person is no longer considered
conforming.
The dl element now represents an
association list of name-value groups, and is no longer said to be
appropriate for dialogue.
The hr element now represents a
paragraph-level thematic break.
The i element now represents a
span of text in an alternate voice or mood, or otherwise offset from the
normal prose in a manner indicating a different quality of text, such as a
taxonomic designation, a technical term, an idiomatic phrase from another
language, a thought, or a ship name in Western texts.
For the label element the
browser should no longer move focus from the label to the control unless
such behavior is standard for the underlying platform user
interface.
The menu element is redefined to
be useful for toolbars and popup menus.
The noscript element is no
longer said to be rendered when the user agent doesn't support a scripting
language invoked by a script element
earlier in the document.
The s element now represents
contents that are no longer accurate or no longer relevant.
The script element can now be
used for scripts or for custom data blocks.
The small element now represents
side comments such as small print.
The strong element now
represents importance rather than strong emphasis.
The u element now represents a
span of text with an unarticulated, though explicitly rendered,
non-textual annotation, such as labeling the text as being a proper name
in Chinese text (a Chinese proper name mark), or labeling the text as
being misspelt.
Several attributes have changed in various ways.
The accept attribute on input now allows the values
audio/*, video/* and image/*.
The accesskey global attribute now allows multiple
characters to be specified, which the user agent can choose from.
The action
attribute on form is no longer allowed
to have an empty URL.
In WHATWG HTML, the method attribute has a new keyword dialog, intended to close a dialog element.
The border attribute on table only allows the values "1" and the
empty string. In WHATWG HTML, the border attribute is obsolete.
The colspan attribute on td and th
now has to be greater than zero.
The coords attribute on area no longer allows a percentage value of
the radius when the element is in the circle state.
The data
attribute on object is no longer said
to be relative to the codebase attribute.
The defer attribute on script now explicitly makes the script
execute when the page has finished parsing.
The dir global
attribute now allows the value auto.
The enctype attribute on form now supports the value
text/plain.
The width
and height
attributes on img, iframe and object are no longer allowed to contain
percentages. They are also not allowed to be used to stretch the image to
a different aspect ratio than its intrinsic aspect ratio.
The href
attribute on link is no longer allowed
to have an empty URL.
The href
attribute on base is now allowed to
contain a relative URL.
All attributes that take URLs, e.g. href on the a element, now support IRIs if the document's
encoding is UTF-8 or UTF-16.
The http-equiv attribute on meta is no longer said to be used by HTTP
servers to create HTTP headers in the HTTP response. Instead, it is said
to be a pragma directive to be used by the user agent.
The id global
attribute is now allowed to have any value, as long as it is unique, is
not the empty string, and does not contain space characters.
The lang global
attribute takes the empty string in addition to a valid language
identifier, just like xml:lang does in XML.
The media attribute on link now accepts a media query and defaults
to "all".
The event handler attributes (e.g. onclick) now always use JavaScript as the
scripting language.
The value
attribute of the li element is no
longer deprecated as it is not presentational. The same goes for the start and type attributes of the
ol element.
The style
global attribute now always uses CSS as the styling language.
The tabindex
global attribute now allows negative values which indicate that the
element can receive focus but cannot be tabbed to.
The target attribute of the a and area
elements is no longer deprecated, as it is useful in Web applications,
e.g. in conjunction with iframe.
The type
attribute on script and style is no longer required if the scripting
language is JavaScript and the styling language is CSS, respectively.
The usemap attribute on img no longer takes a URL, but instead takes
a valid hash-name reference to a map element.
The following attributes are allowed but Web developers are discouraged from using them and instead strongly encouraged to use an alternative solution:
The border
attribute on img. It is required to have
the value "0" when present. Web developers can use CSS
instead.
The language attribute on script. It is required to have the value
"JavaScript" (case-insensitive) when present and cannot
conflict with the type attribute. Web developers can simply omit it as
it has no useful function.
The name attribute on a. Web developers can use the id attribute instead.
The elements in this section are not to be used by Web developers. User
agents will still have to support them and various sections in
HTML define how. E.g. the obsolete isindex element
is handled by the parser section.
The following elements are not in HTML because their effect is purely presentational and their function is better handled by CSS:
The following elements are not in HTML because using them damages usability and accessibility:
The following elements are not included because they have not been used often, created confusion, or their function can be handled by other elements:
acronym is not included because it has created a lot
of confusion. Web developers are to use abbr for
abbreviations.
isindex usage can be replaced by usage of form controls.
Finally the noscript element is only conforming in the
HTML syntax. It is not allowed in the XML syntax. This is because in order to not only hide visually but also prevent the content to run scripts, apply style sheets, have submittable form controls, load resources, and so forth, the HTML parser parses the content of the noscript element as plain text. The same is not possible with an XML parser.
Some attributes from HTML4 are no longer allowed in HTML. The specification defines how user agents should process them in legacy documents, but Web developers are not allowed use them and they will not validate.
HTML has advice on what you can use instead.
In addition, HTML has none of the presentational attributes that were in HTML4 as their functions are better handled by CSS:
align attribute on caption,
iframe, img, input,
object, legend, table,
hr, div, h1, h2,
h3, h4, h5, h6,
p, col, colgroup,
tbody, td, tfoot,
th, thead and tr.
background attribute on body.
cellpadding and cellspacing attributes on
table.
char and charoff attributes on
col, colgroup, tbody,
td, tfoot, th, thead
and tr.
frameborder attribute on iframe.
marginheight and marginwidth attributes on
iframe.
valign attribute on col,
colgroup, tbody, td,
tfoot, th, thead and
tr.
width attribute on hr, table,
td, th, col, colgroup
and pre.
The content model is what defines how elements may be nested — what is allowed as children (or descendants) of a certain element.
At a high level, HTML4 had two major categories of elements, "inline"
(e.g. span, img, text), and "block-level" (e.g. div, hr,
table). Some elements did not fit in
either category.
Some elements allowed "inline" elements (e.g. p), some allowed "block-level" elements (e.g.
body), some allowed both (e.g. div), while other elements did not allow either
category but only allowed other specific elements (e.g. dl, table),
or did not allow any children at all (e.g. link, img,
hr).
Notice the difference between an element itself being in a
certain category, and having a content model of a certain category; for
instance, the p element is itself a
"block-level" element, but has a content model of "inline".
To make it more confusing, HTML4 had different content model rules in its
Strict, Transitional and Frameset flavors; for instance, in Strict, the
body element allowed only "block-level"
elements, but in Transitional, it allowed both "inline" and "block-level".
To make things more confusing still, CSS uses the terms "block-level element" and "inline-level element" for its visual formatting model, which is related to CSS's 'display' property and has nothing to do with HTML's content model rules.
HTML does not use the terms "block-level" or "inline" as part of its content model rules, to reduce confusion with CSS. However, it has more categories than HTML4, and an element can be part of none of them, one of them, or several of them.
Flow content, e.g. span, div, text. This is roughly like HTML4's
"block-level" and "inline" together.
Heading content, e.g. h1.
Phrasing content, e.g. span, img, text. This is roughly like HTML4's
"inline". Elements that are phrasing content are also flow content.
Interactive content, e.g. a, button, label. Interactive content is not allowed to
be nested.
As a broad change from HTML4, HTML no longer has any element that only
accepts what HTML4 called "block-level" elements; e.g. the body element now allows flow content. Thus, This is
closer to HTML4 Transitional than HTML4 Strict.
Further changes include:
The address element now allows
flow content, but with no heading content descendants, no sectioning
content descendants, and no header,
footer, or address element descendants.
WHATWG HTML allows link and
meta as descendants of body if they use microdata attributes.
The noscript element was a
"block-level" element in HTML4, but is phrasing content in HTML.
The table, thead, tbody, tfoot, tr,
ol, ul and dl elements are
allowed to be empty in HTML.
Table elements have to conform to the table model (e.g. two cells are not allowed to overlap).
The table element now does not
allow col elements as direct children.
However, the HTML parser implies a colgroup element, so this change should not
affect text/html content.
The table element now allows the
tfoot element to be the last child.
The caption element now allows
flow content, but with no descendant table elements.
The th element now allows flow
content, but with no header, footer, sectioning content, or heading
content descendants.
The a element now has a transparent content model (except it does not
allow interactive content descendants), meaning that it has the same
content model as its parent. This means that the a element can now contain e.g. div elements, if its parent allows flow
content.
The ins and del elements also have a transparent content
model. HTML4 had similar rules in prose that could not be expressed in the
DTD.
The object element also has a
transparent content model, after its param children.
The map element also has a
transparent content model. The area
element is considered phrasing content if there is a map element ancestor, which means that they
do not need to be direct children of map.
HTML has introduced many new APIs and has extended, changed or obsoleted some existing APIs.
HTML introduces a number of APIs that help in creating Web applications. These can be used together with the new elements introduced for applications:
Media elements (video and audio) have APIs for controlling playback,
syncronising multiple media elements, and timed text tracks (e.g.
subtitles).
An API for form constraint validation (e.g. the setCustomValidity() method).
An API for commands that the user can invoke.
An API that enables offline Web applications, with an application cache.
An API that allows a Web application to register itself for certain
protocols or media types, using registerProtocolHandler()
and registerContentHandler().
Editing API in combination with a new global contenteditable
attribute.
Drag & drop API in combination with a draggable
attribute.
An API that exposes the components of the document's URL and allows
scripts to navigate, redirect and reload (the Location interface).
An API that exposes the session history and allows scripts to
update the document's URL without actually navigating, so that
applications don't need to abuse the fragment component for "Ajax-style"
navigation (the History interface).
An API to schedule timer-based callbacks (setTimeout() and setInterval()).
An API to prompt the user (alert(), confirm(), prompt(), showModalDialog()).
An API for printing the document (print()).
An API for handling search providers (AddSearchProvider() and IsSearchProviderInstalled()).
The Window object has been defined.
WHATWG HTML has further APIs that are not in W3C HTML5 but are separate specifications at the W3C:
An API for microdata.
An API for immediate-mode bitmap graphics (the 2d context for the
canvas element).
An API for cross-document messaging and channel messaging (postMessage()
and MessageChannel).
An API for runnings scripts in the background (Worker and SharedWorker).
An API for client-side storage (localStorage and sessionStorage).
An API for bidirectional client-server communication (WebSocket).
An API for server-to-client data push (EventSource).
The following features from DOM Level 2 HTML are changed in various ways:
document.title now collapses whitespace on
getting.
document.domain is made settable, which
can change the document's effective script origin.
document.open() now either clears the
document (if invoked with two or less arguments), or acts like window.open() (if invoked with
three or four arguments). In the former case, throws an exception in XML.
document.close(), document.write() and
document.writeln() throw an exception in
XML. The latter two now support variadic arguments; they can add text to
the document's input stream while it is still being parsed, imply a
call to document.open(), or be ignored altogether in
some cases.
document.getElementsByName()
now returns all HTML elements with a name attribute matching
the argument.
elements
on HTMLFormElement now returns an
HTMLFormControlsCollection of button, fieldset, input, keygen, object, output, select and textarea elements. length returns the
number of nodes in elements.
add() on
HTMLSelectElement now also accepts an
integer as its second argument.
remove()
on HTMLSelectElement now removes the
first element in the collection if the argument is out of bounds.
The click(),
focus() and blur() methods are now
available on all HTML elements.
DocumentDOM Level 2 HTML had an HTMLDocument interface that
inherited from Document and provided HTML-specific members on
documents. HTML has moved these members to the Document interface, and extended it in a number
of ways. Since all documents use the Document interface, the
HTML-specific members are now available on all documents, so they are usable
in e.g. SVG documents as well. It also has several new members:
location, lastModified and readyState to
help resource metadata management.
dir,
head, embeds, plugins, scripts, commands, and a
generic name getter, to access various parts of the DOM tree.
WHATWG HTML has getItems() for microdata.
WHATWG HTML and W3C HTML5.1 have cssElementMap to accompany the CSS
element() feature.
WHATWG HTML and W3C HTML5.1 have currentScript which returns the currently
running script element (or null).
activeElement and hasFocus to
determine which element is currently focused and whether the Document has focus respectively.
designMode, execCommand(),
queryCommandEnabled(), queryCommandIndeterm(),
queryCommandState(), queryCommandSupported(), queryCommandValue() for the
editing API.
All event handler IDL attributes. Also, onreadystatechange is a
special event handler IDL attribute that is only available on
Document.
Existing scripts that modified the prototype of HTMLDocument
should continue to work because window.HTMLDocument now returns
the Document interface object.
HTMLElementThe HTMLElement interface has also gained several extensions
in HTML:
translate,
hidden, tabIndex, accessKey, draggable, dropzone, contentEditable,
contextMenu,
spellcheck and
style reflect content
attributes.
dataset is a
convenience feature for handling the data-* attributes, which are exposed as
camel-cased properties; for instance, elm.dataset.fooBar
= 'test' sets the data-foo-bar content attribute on
elm.
WHATWG HTML has itemScope, itemType, itemId, itemRef, itemProp, properties and itemValue for microdata.
click(), focus() and blur() allow scripts to
simulate clicks and moving focus.
accessKeyLabel gives the shortcut key that
the user agent has assigned for the element, which the Web developer can
influence with the accesskey attribute.
isContentEditable returns true if the
element is editable.
forceSpellCheck() causes the user agent to check spelling of an element.
commandType, commandLabel, commandIcon, commandHidden, commandDisabled and commandChecked is
part of the command API.
All event handler IDL attributes.
Some members were previously defined on HTMLElement but been moved to the Element
interface in the DOM standard: [DOM]
id reflects the id content attribute.
className reflects the class content attribute.
classList is a convenient accessor for
className. The object it returns exposes methods (contains(), add(),
remove(), and toggle()) for manipulating the element's classes.
getElementsByClassName() returns a list of elements with the specified
classes.
Some interfaces in DOM Level 2 HTML have been extended.
HTMLOptionsCollection now has a
legacy caller, setter creator, and the members add(), remove() and selectedIndex
HTMLLinkElement and HTMLStyleElement now implement the
LinkStyle interface from CSSOM. [CSSOM]
HTMLFormElement now has a named
getter and an indexed getter.
HTMLSelectElement now has a
getter, item()
and namedItem() methods, a setter creator,
selectedOptions and labels IDL attributes,
and members for the form constrain validation API: willValidate,
validity, validationMessage, checkValidity()
and setCustomValidity().
HTMLOptionElement now has a
constructor Option.
HTMLInputElement now has the
members files,
height, indeterminate,
list, valueAsDate,
valueAsNumber, width, stepUp(), stepDown(), the form
constraint validation API members, labels, and members for the text field selection
API: selectionStart, selectionEnd, selectionDirection,
setSelectionRange()
and setRangeText().
HTMLTextAreaElement now has the
members textLength, the form constraint
validation API members, labels and the text field selection API
members.
HTMLButtonElement now has the
form constraint validation API members and labels.
HTMLLabelElement now has the
member control.
HTMLFieldSetElement now has the
members type,
elements
and the form constraint validation API members.
HTMLAnchorElement now has the
members relList,
text, and implements the URLUtils interface which has the members
href, origin,
protocol, username,
password, host,
hostname, port,
pathname, search,
query and hash.
HTMLLinkElement and HTMLAreaElement also have the relList IDL attribute.
HTMLAreaElement also implements the URLUtils interface.
HTMLImageElement now has a
constructor Image, and the
members naturalWidth, naturalHeight and
complete.
HTMLObjectElement now has the
members contentWindow, the form constraint
validation API members and a legacy caller.
HTMLMapElement now has the
member images.
HTMLTableElement now has the
members createTBody() and, in WHATWG HTML and W3C HTML5.1,
stopSorting().
In WHATWG HTML and W3C HTML5.1,
HTMLTableHeaderCellElement now has the member
sort().
HTMLIFrameElement now has the
member contentWindow.
In addition, most new content attributes also have corresponding IDL
attributes on the elements' interfaces, e.g. the sizes IDL attribute on HTMLLinkElement which reflects the sizes content attribute.
Some APIs are now either removed altogether, or marked as obsolete.
All IDL attributes that reflect a content attribute that is itself
obsolete, are now also obsolete; for instance, the bgColor IDL attribute on
HTMLBodyElement which reflects the
obsolete bgcolor
content attribute is now obsolete.
The following interfaces are marked obsolete since the elements are
obsolete: HTMLAppletElement, HTMLFrameSetElement, HTMLFrameElement, HTMLDirectoryElement and HTMLFontElement.
The HTMLIsIndexElement interface is removed altogether since
the HTML parser expands an isindex tag into other elements. The
HTMLBaseFontElement interface is also removed since the element has no effect.
The following members of the HTMLDocument interface (which
have now moved to Document) are now
obsolete: anchors and applets.
The changelogs in this section indicate what has been changed in WHATWG HTML, W3C HTML5 and W3C HTML5.1 between somewhat arbitrary dates, often close to publications of the W3C HTML5 or W3C HTML5.1 drafts. Rationale for changes can be found in the public-html@w3.org and whatwg@whatwg.org mailing list archives, and the WHATWG Weekly series of blog posts. More fundamental rationale is being collected on the WHATWG Rationale wiki page. Many editorial and minor technical changes are not included in these changelogs. Implementors are strongly encouraged to follow the development of the main specification on a frequent basis so they become aware of all changes that affect them early on.
The changes in the changelogs are in rough reverse chronological order.
The canvas 2d context has a new method isPointInStroke(). (WHATWG HTML and W3C HTML5.1.)
The DataTransferItemList interface (part of drag and drop API) now has a remove() method instead of a deleter. (WHATWG HTML and W3C HTML5.1.)
The registerProtocolHandler() API now supports bitcoin: and geo: schemes. (bitcoin: is WHATWG HTML only.)
The alert(), confirm() and prompt() methods now have all arguments optional. (WHATWG HTML and W3C HTML5.1.)
The fieldset element can now match the :invalid pseudo-class. (WHATWG HTML and W3C HTML5.1.)
The HTMLBaseFontElement interface has been removed. (WHATWG HTML and W3C HTML5.1.)
The specification has integrated with the URL standard which effectively added the origin, username, password and query IDL attributes to the a and area elements and the location object. (WHATWG HTML and W3C HTML5.1.)
The hgroup element has been dropped. (W3C HTML5 and W3C HTML5.1.)
The HTML parser now better supports innerHTML on non-HTML elements. (WHATWG HTML and W3C HTML5.1.)
Various things that were present in WHATWG HTML but missing in W3C HTML5.1 were "un-forked": The download attribute on a and area, the application cache prefer-online feature, the text range API on input and textarea, the cssElementMap IDL attribute on Document, various bits for microdata. (W3C HTML5.1 only.)
The drag and drop API now has a dragexit event. (WHATWG HTML and W3C HTML5.1.)
The form element now recieves an invalid event when submission fails due to form validation (in addition to the individual controls). (WHATWG HTML and W3C HTML5.1.)
The navigator object now has a product IDL attribute that always returns "Gecko" (for compatibility). (WHATWG HTML and W3C HTML5.1.)
The time element's datetime IDL attribute got renamed to dateTime. (WHATWG HTML and W3C HTML5.1.)
The pushState() and replaceState() methods can now be invoked with null as the third argument.
A new main element has been added.
The stepUp() and stepDown() methods on the input element have been tweaked. (WHATWG HTML and W3C HTML5.1.)
The img element now supports progress events. (WHATWG HTML and W3C HTML5.1.)
A new fillRule IDL attribute has been added to the canvas 2d context. (WHATWG HTML only.)
The style IDL attribute on HTML elements is now settable. (WHATWG HTML and W3C HTML5.1.)
The media attribute has been dropped from the a and area elements. (WHATWG HTML and W3C HTML5.1.)
Heading elements now support automatic sizing in hgroup if they follow an h1 element. (WHATWG HTML only.)
The navigator object now has a language IDL attribute. (WHATWG HTML and W3C HTML5.1.)
The menu API has been revamped. The command element was dropped. An new menuitem element was added. The menu element's type attribute now uses the value "popup" instead of "context". The button element has a new menu attribute and the type attribute supports a new value "menu". (WHATWG HTML and W3C HTML5.1.)
The table element now supports sorting columns. The table element has a new sortable attribute and a stopSorting() method. The th element has a new sorted attribute and a sort() method. (WHATWG HTML and W3C HTML5.1.)
The autocomplete attribute on form controls now supports the cc-type type. (WHATWG HTML and W3C HTML5.1.)
A new currentScript IDL attribute on document was added. (WHATWG HTML and W3C HTML5.1.)
The link and script elements now support the crossorigin attribute. (WHATWG HTML and W3C HTML5.1.)
The canvas element now supports indirect and proxied rendering contexts, to support drawing from a worker. The CanvasProxy, ImageBitmap interfaces are introduced, the canvas element has new setContext() and transferControlToProxy() methods, and a new createImageBitmap() method on window was introduced. (WHATWG HTML and W3C HTML5.1.)
The canvas 2d context has a new direction IDL attribute. (WHATWG HTML only.)
The screen object has a new canvasResolution IDL attribute. (WHATWG HTML only.)
A new PortCollection() constructor was added to support sending messages to many ports while allowing them to be garbage collected. (WHATWG HTML only.)
The getItem() method on Storage can now return null. (WHATWG HTML only.)
The sandbox attribute on iframe supports a new value "allow-pointer-lock". (WHATWG HTML and W3C HTML5.1.)
The HTML parser now invokes mutation observers. (WHATWG HTML and W3C HTML5.1.)
The canvas element has a new supportsContext() method. (WHATWG HTML and W3C HTML5.1.)
The abbr attribute on the th element is now conforming.
A new forceSpellCheck() method on HTML elements was added. (WHATWG HTML and W3C HTML5.1.)
The EventSource API now reconnects for DNS errors and TCP-level connection failures. (WHATWG HTML only.)
A new fastSeek() method was added to media elements. (WHATWG HTML and W3C HTML5.1.)
The http+aes: and https+aes: schemes were removed due to lack of interest. (WHATWG HTML only.)
Navigations with rel=noreferrer now does not clone the sessionStorage.
The error event for Worker objects now has a column IDL attribute. (WHATWG HTML only.)
The autofocus attribute on form controls in dialog elements now has an effect when the showModal() method is invoked.
The content model for ruby was changed with regards to nested ruby elements.
Self-closing SVG script tags in the HTML syntax now execute.
The placeholder section for the find() API has been dropped.
An encoding declaration is now required in the HTML syntax even if only ASCII characters are used.
Some bug fixes in the Drag and Drop API.
The inBandMetadataTrackDispatchType IDL attribute was added to TextTrack.
The TextTrackCue() constructor now has fewer arguments.
The accept attribute now supports file extensions as well as MIME types.
The initialTime IDL attribute on media elements has been dropped.
The startOffsetTime IDL attribute on media elements has been renamed to startDate.
Further changes to WHATWG HTML that do not affect W3C HTML5:
Several changes and bug fixes in the Text Track API.
addElement() was dropped from the Drag and Drop API.
Media queries are now proxied for iframe elements with the seamless attribute.
The :enabled and :disabled pseudo-classes now apply to input elements in the Hidden state.
The ssh, sip and magnet schemes are now in the registerProtocolHandler() whitelist.
table elements now have 'box-sizing:border-box' by default.
Bug fixes in the "potentially CORS-enabled fetch" algorithm.
The document outline algorithm now ignores elements with the hidden attribute.
Markup generators that are unable to provide required alt text can now use a specific attribute on img that makes validators ignore the missing alt error.
Workers and shared workers now support data: URLs.
The inputmode attribute has been added to input and textarea.
The autocomplete attribute has been extended to support prefilling specific things.
WebSocket supports sending ArrayBufferView as well as ArrayBuffer.
The border attribute on table is non-conforming again.
The canvas ImageData methods now assume 96dpi, and a set of "HD" methods have been introduced.
The shared worker connect event now also exposes the source port in the source IDL attribute.
Lone surrogates are converted to U+FFFD instead of throwing in WebSocket send().
The setRangeText() method has been added to input and textarea.
The srcset attribute has been added to img.
Application cache now has an prefer-online mode.
Dialogs are now supported with the dialog element, the inert global attribute and the dialog method on form.
The resetTransform() method, currentTransform IDL attribute, several IDL attributes for font metrics, resetClip() method, imageSmoothingEnabled IDL attribute, addHitRegion() method, removeHitRegion() method, support for dashed lines, have been added to the canvas 2d context.
Support for mutation observers was added.
The TextTrackCue members alignment, linePosition, textPosition and direction were renamed to align, line, position and vertical, respectively.
The command element now has a command attribute.
Drag and drop content is now suggested to be filtered by user agents to prevent XSS attacks.
The translate global attribute was added.
The showModalDialog(), alert(), confirm() and prompt() methods are now allowed to do nothing during pagehide, beforeunload and unload events.
The script element now supports beforescriptexecute and afterscriptexecute events.
window.onerror now supports a fourth argument for column position.
The window.opener IDL attribute can now return null in some cases.
The clearTimeout() and clearInterval() methods were made synonymous.
The CSS @global at-rule was introduced, for use together with style elements with the scoped attribute.
The embed and object elements now have a legacy caller.
The handling of window.onerror's return value was changed to match reality.
The setTimeout() API is now allowed to be throttled in background tabs.
The :valid and :invalid pseudo-classes now apply to form elements.
The toBlob() method on canvas now honors the origin-clean flag.
The activeElement IDL attribute now points to the relevant browsing context container (e.g. iframe) when a child document has focus.
The atob() method now ignores whitespace.
The dropzone attribute was changed to use "string:" and "file:" instead of "s:" and "f:".
The HTML parser was fixed to correctly handle a case involving foreign lands and foster parenting.
The date-and-time microsyntaxes now allows a single space instead of a "T".
Application cache no longer checks the MIME type of the cache manifest.
The cueAsSource IDL attribute on TextTrackCue got renamed to text.
The window.onerror API is now invoked with dummy arguments for cross-origin scripts.
The textarea element's value and textLength IDL attributes have their newlines normalized to LF.
The q element now has language-specific quotes rendered by default.
The data element was introduced.
The time element was redesigned to make it match how people wanted to use it. Its pubdate attribute was dropped.
The legacy caller on form was removed.
The location.resolveURL() method was removed.
The track element now sniffs instead of obeying the MIME type.
The load() method on documents created by createDocument() is now defined on the XMLDocument interface.
Members of HTMLDocument moved to Document and window.HTMLDocument now just returns window.Document.
The MutableTextTrack and TextTrack interfaces were merged and TextTrackCue was made more mutable.
The selectedOption IDL attribute on input was dropped.
Attribute values in Selectors are now case sensitive for all attributes.
The readyState IDL attribute moved from TextTrack to HTMLTrackElement.
The text/html-sandboxed MIME type was dropped.
Floating point numbers are now allowed to begin with a "." character.
Navigating to an audio or video resource is now supported.
Table cells now allow flow content but does not allow header, footer, sectioning content or heading content descendants.
Adding a track to a media element now fires an addtrack event on the relevant track list objects.
Setting currentTime on media elements before the media has loaded now defers the seek instead of throwing.
Plugins are no longer disabled in sandboxed iframes if they honor the sandbox attribute.
Some tweaks to history navigation and related events.
Media elements and MediaControllers now get paused when they end.
Events now support constructors and some init*Event() methods were removed.
Media elements now fire a suspend event when the resource is loaded.
Form submission now normalizes newlines to CRLF.
Some tweaks around bidi and the br element.
Large parts of the Editing section moved to HTML Editing APIs.
UndoManager and related features moved to UndoManager and DOM Transaction.
isProtocolHandlerRegistered(), isContentHandlerRegistered(), unregisterProtocolHandler() and unregisterContentHandler() were added.
registerContentHandler() now has a blacklist of MIME types.
registerProtocolHandler() now has a whitelist of protocols, but also supports any protocol that starts with "web+".
Fragment identifiers for text/html resources now don't need to point to an element with a matching ID.
audio elements are now allowed to have zero source children.
There are now some restrictions on the use of bidi formatting characters.
The maxlength and size attributes are allowed (but give warnings in validators) on input elements with type=number.
The link relation "shortcut icon" is now allowed.
Heading elements are now allowed to have the heading and tab roles.
Things that use EventTarget now inherit from it instead of using "implements".
The setInterval() API now clamps to 4ms instead of 10ms.
The select element and its options collection now have a setter.
rel=help on links now show a help cursor by default.
Calls to window.print() before the document is loaded defers the print until it is loaded.
Application cache gained an abort() method.
HTMLCollection, DOMTokenList, getElementsByClassName(), createHTMLDocument(), HTML-specific overrides to some DOM Core features (like createElement()), some definitions, the id IDL attribute and ID handling moved to DOM4.
Fragment identifiers can now survive redirects.
The pushState() and replaceState() methods now change the history entry to GET.
The command API now has its properties prefixed so they are now commandLabel, commandIcon, commandHidden, commandDisabled and commandChecked.
The structured clone algorithm now supports sparse arrays.
window.postMessage now supports transferring some objects instead of cloning them, and supports transferring ArrayBuffer.
Application cache was made stricter in its MIME type checking.
The placeholder attribute is now allowed on input elements with type=number.
MediaController gained an onended event listener.
The HTML parser changed its handling of U+0000 characters in some places.
The object element gained a new attribute typemustmatch, to make it safer for Web developers to embed untrusted resources where they expect a certain content type.
The form attribute was removed from meter and progress.
The HTML parser was made more forward compatible in its handling of ruby.
Some MIME types (e.g. text/plain) that are guaranteed to never be supported as scripting types for script were specified, so Web developers can safely use them for custom data blocks.
about:blank documents created from window.open() now get a load event.
window.status was specified to exist but do nothing.
Drag and drop DataTransferItems was renamed to DataTransferItemList.
Application cache now supports 'no-store' and HTTPS.
The structured clone algorithm now supports getters.
The crossorigin attribute has been added to img, video and audio to use CORS.
The external IDL attribute has been added on window and has the members AddSearchProvider() and IsSearchProviderInstalled().
Further changes to WHATWG HTML that do not affect W3C HTML5:
The 2d context now supports ellipses with the arc() and arcTo() methods and the new ellipse() method.
The 2d context now supports Path objects. SVG path data can be added to a Path.
The http+aes: and https+aes: URL schemes were added to allow sensitive resources to be held on untrusted servers.
When the itemprop attribute is used on an element where microdata gets its value from an attribute (like href on a elements), that attribute is now required.
PeerConnection was moved to WebRTC.
WebVTT was moved to its own specification.
WebSockets no longer receive messages in the CLOSING state.
The Atom conversion algorithm was dropped.
The itemtype attribute now allows multiple types.
CanvasPixelArray was dropped in favor of Uint8ClampedArray.
The microdata to RDF conversion algorithm was dropped.
The link element is no longer allowed to have both rel and itemprop.
WebSocket API disallows opening an insecure connection if the document uses a secure connection.
The "storage mutex" is made optional.
Web Storage no longer supports structured data.
The a element got a new download attribute. This attribute is not included in W3C HTML5.
An experimental specification for the window.find() method was added.
The 2d context fillText() and strokeText() methods now do not collapse whitespace.
Microdata now handles infinite loops.
Web Worker location now stringifies.
Script errors in a Web Worker can now be detected in a parent worker or the document with the onerror handler.
EventSource now supports CORS.
EventSource was made stricter in its MIME type checking.
Web Workers gained the atob() and btoa() methods.
Web Workers gained the ononline and onoffline event handlers.
WebSockets API has the error event again.
WebSockets API now exposes the selected extensions.
Various tweaks to the UDP PeerConnection API.
WebSocket close code and reason are now supported in the API.
Binary data is now supported in WebSockets.
Redirects in WebSockets are now blocked for security reasons.
Support for the javascript: scheme in img, object, CSS, etc, has been dropped.
The toBlob() method has been added to canvas.
The drawFocusRing() method on the canvas 2d context has been split into two methods, drawSystemFocusRing() and drawCustomFocusRing().
The values attribute on PropertyNodeList has been replaced with a getValues() method.
The select event has been specified.
The selectDirection IDL attribute has been added to input and textarea.
The :enabled and :disabled pseudo-classes now match fieldset, and the :indeterminate pseudo-class can now match progress.
The getKind() method has been added to TrackList.
The MediaController API and the mediagroup attribute have been added to synchronize playback of media elements.
Some ARIA defaults have changed, and it is now invalid to specify ARIA attributes that match the defaults.
The getName() method on TrackList was renamed to getLabel().
The border attribute on table is now conforming.
The u element is now conforming.
The summary attribute on table is now non-conforming.
The audio attribute on video was changed to a boolean muted attribute.
The Content-Language meta pragma is now non-conforming.
The pushState and replaceState features have
been changed based on implementation feedback in Firefox, and history.state
has been introduced.
The tracks IDL attribute on media elements has been renamed
to textTracks.
Event handler content attributes now support JavaScript strict mode.
The forminput and formchange events, and
the dispatchFormInput() and dispatchFormChange()
methods have been dropped.
The rel keywords archives, up,
last, index, first and related synonyms
have been dropped.
Removing a media element from the DOM and inserting it again in the same script now doesn't pause the media element.
The video element's letterboxing rules are now specified in
terms of CSS 'object-fit'.
Cross-origin fonts now don't leak information about the font when drawn
on a canvas.
The character encoding declaration is now allowed to be within the first 1024 bytes instead of the first 512 bytes.
The onerror event handler on window is now invoked
for compile-time script errors as well as runtime errors.
Script-inserted script elements now have async
default to true, which can be set to false to make
the scripts execute in insertion order.
The atob() and btoa() methods have been specified.
The suggested file extension for application cache manifest files has been
changed from .manifest to .appcache.
The action and formaction attributes are
no longer allowed to have the empty string as value.
Drag and drop model was refined.
A new global dropzone attribute was added.
A new bdi element was added to aid with user-generated
content that may have bidi implications.
The dir attribute gained a new "auto"
value.
A dirname attribute was added to input
elements. When specified the directionality as specified by the user
will be submitted to the server as well.
A new track element and associated TextTrack API were added for video text tracks.
The type attribute on the ol element is now allowed.
The getSelection() API moved to a separate
DOM Range draft.
Similarly UndoManager has been removed from the W3C copy of
HTML5 for now as it is not ready yet.
Numerous changes to the HTML parsing algorithm based on implementation feedback.
The hidden attribute now works for table-related
elements.
The canvas getContext() method is now
defined to be able to handle multiple contexts better.
The media elements' startTime IDL attribute was
renamed to initialTime and startOffsetTime
was added.
The prefetch link relationship can now be used on
a elements.
The datetime attribute of ins and
del no longer requires a time to be specified.
Using PUT and DELETE as HTTP methods for the form
element is no longer supported.
The s element is no longer deprecated.
The video element has a new audio
attribute.
Per usual, lots of other minor fixes have been made as well.
The ping attribute has been removed from the W3C
version of HTML5.
The title element is optional for
iframe srcdoc documents and other scenarios
where a title is already available. As is the case with email.
keywords is now a standard metadata name for the
meta element.
The allow-top-navigation value has been added for the
sandbox attribute on the iframe element. It
allows the embedded content to navigate its parent when specified.
The wbr element has been added.
The alternate keyword for the rel
attribute of the link element can now be used to point to
feeds again, even if the feed is not an alternative for the
document.
The HTML to Atom mapping has been removed from the W3C version of HTML5.
In addition lots of minor changes, clarifications, and fixes have been made to the document.
The dialog element has been removed. A section with
advice on how to mark up conversations has effectively replaced
it.
document.head has been introduced to provide
convenient access to the head element from script.
The link type feed has been removed.
alternate with specific media types
is to be used instead.
createHTMLDocument() has been introduced as API to
allow easy creation of HTML documents.
Both the meter and progress elements no
longer have "magic" processing of their contents because it could not
be made to work internationally.
The meter and progress elements, as well
as the output element, can now be labeled using the
label element.
A new media type, text/html-sandboxed, was introduced
to allow hosting of potentially hostile content without it causing
harm.
A srcdoc attribute for the iframe element
was introduced to allow embedding of potentially hostile content
inline. It is expected to be used together with the
sandbox and seamless attributes.
The figure element now uses a new element
figcaption rather than legend because people
want to use HTML long before it reaches W3C Recommendation.
The details element now uses a new element
summary for exactly the same reason.
The autobuffer attribute on media elements was renamed
to preload.
A whole lot of other smaller issues have also been resolved. The above list summarizes what is thought to be of primary interest to Web developers.
In addition to all of the above, Microdata, the 2D context API for
canvas, and Web Messaging (postMessage() API)
have been split into their own drafts at the W3C (the WHATWG still
publishes a version of HTML that includes them):
Specific microdata vocabularies are gone altogether in the W3C draft of HTML5 and are not published as a separate draft. The WHATWG draft of HTML still includes them.
When the time element is empty user agents have to
render the time in a locale-specific manner.
The load event is dispatched at Window,
but now has Document as its target.
pushState() now affects the Referer (sic)
header.
onundo and onredo are now on
Window.
Media elements now have a startTime member that
indicates where the current resource starts.
header has been renamed to hgroup and a
new header element has been introduced.
createImageData() now also takes
ImageData objects.
createPattern() can now take a video
element as argument too.
The footer element is no longer allowed in
header and header is not allowed in
address or footer.
A new control has been introduced:
<input type="tel">
The Command API now works for all elements.
accesskey is now properly defined.
section and article now take a
cite attribute.
A new feature called Microdata has been introduced which allows people to embed custom data structures in their HTML documents.
Using the Microdata model three predefined vocabularies have also been included: vCard, vEvent, and a model for licensing.
Drag and drop has been updated to work with the Microdata model.
The last of the parsing quirks has been defined.
textLength has been added as member of the
textarea element.
The rp element now takes phrasing content rather than
a single character.
location.reload() is now defined.
The hashchange event now fires asynchronously.
Rules for compatibility with XPath 1.0 and XSLT 1.0 have been added.
The spellcheck IDL attribute now maps to a
DOMString.
hasFeature() support has been reduced to a
minimum.
The Audio() constructor sets the
autobuffer attribute.
The td element is no longer allowed in
thead.
The input element and DataTransfer object
now have a files IDL attribute.
The datagrid and bb have been removed due
to their design not being agreed upon.
The cue range API has been removed from the media elements.
Support for WAI-ARIA has been integrated.
On top of this list quite a few minor clarifications, typos, issues specific to implementors, and other small problems have been resolved.
In addition, the following parts of HTML have been taken out and will likely be further developed at the IETF:
Definition of URLs.
Definition of Content-Type sniffing.
A new global attribute called spellcheck has been
added.
Defined that JavaScript this in the global object
returns a WindowProxy object rather than the
Window object.
The value IDL attribute for input
elements in the File Upload state is now defined.
Definition of designMode was changed to be more in
line with legacy implementations.
The drawImage() method of the 2D drawing API can now
take a video element as well.
The way media elements load resources has been changed.
document.domain is now IPv6-compatible.
The video element gained an autobuffer
boolean attribute that serves as a hint.
You are now allowed to specify the meta element with a
charset attribute in XML documents if the value of that
attribute matches the encoding of the document. (Note that it does not
specify the value, it is just a talisman.)
The bufferingRate and bufferingThrottled
members of media elements have been removed.
The media element resource selection algorithm is now asynchronous.
The postMessage() API now takes an array of
MessagePort objects rather than just one.
The second argument of the add() method on the
select element and the options member of the
select element is now optional.
The action, enctype, method,
novalidate, and target attributes on
input and button elements have been
renamed to formaction, formenctype,
formmethod, formnovalidate, and
formtarget.
A "storage mutex" concept has been added to deal with separate
pages trying to change a storage object (document.cookie
and localStorage) at the same time. The
Navigator gained a getStorageUpdates() method
to allow it to be explicitly released.
A syntax for SVG similar to MathML is now defined so that SVG can
be included in text/html resources.
The placeholder attribute has been added to
the textarea element.
Added a keygen element for key pair generation.
The datagrid element was revised to make the API more
asynchronous and allow for unloaded parts of the grid.
In addition, several parts of HTML have been taken out and will be further developed by the Web Applications Working Group as standalone specifications:
Web Storage (localStorage and sessionStorage)
The data member of ImageData objects has
been changed from an array to a CanvasPixelArray
object.
Shadows are now required from implementations of the
canvas element and its API.
Security model for canvas is clarified.
Various changes to the processing model of canvas have
been made in response to implementation and Web developer feedback. E.g.
clarifying what happens when NaN and Infinity are passed and fixing the
definitions of arc() and arcTo().
innerHTML in XML was slightly changed to improve
round-tripping.
The toDataURL() method on the canvas
element now supports setting a quality level when the media type
argument is image/jpeg.
The poster attribute of the video element
now affects its intrinsic dimensions.
The behavior of the type attribute of the
link element has been clarified.
Sniffing is now allowed for link when the expected
type is an image.
A section on URLs is introduced dealing with how URL values are to be interpreted and what exactly Web developers are required to do. Every feature of the specification that uses URLs has been reworded to take the new URL section into account.
It is now explicit that the href attribute of the
base element does not depend on
xml:base.
It is now defined what the behavior should be when the base URL changes.
URL decomposition IDL attributes are now more aligned with Internet Explorer.
The xmlns attribute with the value
http://www.w3.org/1999/xhtml is now allowed on all HTML
elements.
data-* attributes and custom attributes on
the embed element now have to match the XML
Name production and cannot contain a colon.
WebSocket API is introduced for bidirectional communication with a server.
The default value of volume on media elements is now
1.0 rather than 0.5.
event-source was renamed to eventsource
because no other HTML element uses a hyphen.
A message channel API has been introduced augmenting
postMessage().
A new element named bb has been added. It represents a
user agent command that the user can invoke.
The addCueRange() method on media elements has been
modified to take an identifier which is exposed in the callbacks.
It is now defined how to mutate a DOM into an infoset.
The parent attribute of the Window object
is now defined.
The embed element is defined to do extension sniffing
for compatibility with servers that deliver Flash as
text/plain. (This is marked as an issue in the
specification to figure out if there is a better way to make this
work.)
The embed can now be used without its src
attribute.
getElementsByClassName() is defined to be ASCII
case-insensitive in quirks mode for consistency with CSS.
In HTML documents localName no longer returns the node
name in uppercase.
data-* attributes are defined to be always
lowercase.
The opener attribute of the Window object
is not to be present when the page was opened from a link with
target="_blank" and rel="noreferrer".
The top attribute of the Window object is
now defined.
The a element now allows nested flow content, but not
nested interactive content.
It is now defined what the header element means to
document summaries and table of contents.
What it means to fetch a resource is now defined.
Patterns are now required for the canvas element.
The autosubmit attribute has been removed from the
menu element.
Support for outerHTML and
insertAdjacentHTML() has been added.
xml:lang is now allowed in HTML when lang
is also specified and they have the same value. In XML
lang is allowed if xml:lang is also specified
and they have the same value.
The frameElement attribute of the Window
object is now defined.
An event loop and task queue is now defined detailing script execution and events. All features have been updated to be defined in terms of this mechanism.
If the alt attribute is omitted a title
attribute, an enclosing figure element with a
legend element descendant, or an
enclosing section with an associated heading must be present.
The irrelevant attribute has been renamed to
hidden.
The definitionURL attribute of MathML is now properly
supported. Previously it would have ended up being all lowercase during
parsing.
User agents must treat US-ASCII as Windows-1252 for compatibility reasons.
An alternative syntax for the DOCTYPE is allowed for compatibility with some XML tools.
Data templates have been removed (consisted of the
datatemplate, rule and nest
elements).
The media elements now support just a single loop
attribute.
The load() method on media elements has been
redefined as asynchronous. It also tries out files in turn now rather
than just looking at the type attribute of the
source element.
A new member called canPlayType() has been added to
the media elements.
The totalBytes and bufferedBytes
attributes have been removed from the media elements.
The Location object gained a
resolveURL() method.
The q element has changed again. Punctuation is to be
provided by the user agent again.
Various changes were made to the HTML parser algorithm to be more in line with the behavior Web sites require.
The unload and beforeunload events are
now defined.
The IDL blocks in the specification have been revamped to be in line with the upcoming Web IDL specification.
Table headers can now have headers. User agents are required to
support a headers attribute pointing to a td
or th element, but Web developers are required to only let them
point to th elements.
Interested parties can now register new http-equiv
values.
When the meta element has a charset
attribute it must occur within the first 512 bytes.
The StorageEvent object now has a
storageArea attribute.
It is now defined how HTML is to be used within the SVG
foreignObject element.
The notification API has been dropped.
How [[Get]] works for the HTMLDocument and
Window objects is now defined.
The Window object gained the
locationbar, menubar,
personalbar, scrollbars,
statusbar and toolbar attributes giving
information about the user interface.
The application cache section has been significantly revised and updated.
document.domain now relies on the Public Suffix List.
[PSL]
A non-normative rendering section has been added that describes user agent rendering rules for both obsolete and conforming elements.
A normative section has been added that defines when certain selectors as defined in the Selectors and the CSS3 Basic User Interface Module match HTML elements. [SELECTORS] [CSSUI]
Web Forms 2.0, previously a standalone specification, has been fully integrated into HTML since last publication. The following changes were made to the forms chapter:
Support for XML submission has been removed.
Support for form filling has been removed.
Support for filling of the select and
datalist elements through the data attribute
has been removed.
Support for associating a field with multiple forms has been
removed. A field can still be associated with a form it is not nested
in through the form attribute.
The dispatchChangeInput() and
dispatchFormChange() methods have been removed from the
select, input, textarea, and
button elements.
Repetition templates have been removed.
The inputmode attribute has been removed.
The input element in the File Upload state no longer
supports the min and max attributes.
The allow attribute on input elements in
the File Upload state is no longer authoritative.
The pattern and accept attributes for
textarea have been removed.
RFC 3106 is no longer explicitly supported.
The submit() method now just submits, it no longer
ensures the form controls are valid.
The input element in the Range state now defaults to
the middle, rather than the minimum value.
The size attribute on the input element
is now conforming (rather than deprecated).
object elements now partake in form submission.
The type attribute of the input element
gained the values color and search.
The input element gained a multiple
attribute which allows for either multiple e-mails or multiple files to
be uploaded depending on the value of the type
attribute.
The input, button and form
elements now have a novalidate attribute to indicate that
the form fields should not be required to have valid values upon
submission.
When the label element contains an input
it may still have a for attribute as long as it points to
the input element it contains.
The input element now has an
indeterminate IDL attribute.
The input element gained a placeholder
attribute.
Implementation and authoring details around the ping
attribute have changed.
<meta http-equiv=content-type> is now a
conforming way to set the character encoding.
API for the canvas element has been cleaned up. Text
support has been added.
globalStorage is now restricted to the same-origin
policy and renamed to localStorage. Related event
dispatching has been clarified.
postMessage() API changed. Only the origin of the
message is exposed, no longer the URL. It also requires a second
argument that indicates the origin of the target document.
Drag and drop API has got clarification. The
dataTransfer object now has a types
attribute indicating the type of data being transferred.
The m element is now called mark.
Server-sent events has changed and gotten clarification. It uses a new format so that older implementations are not broken.
The figure element no longer requires a caption.
The ol element has a new reversed
attribute.
Character encoding detection has changed in response to feedback.
Various changes have been made to the HTML parser section in response to implementation feedback.
Various changes to the editing section have been made, including
adding queryCommandEnabled() and related methods.
The headers attribute has been added for
td elements.
The table element has a new createTBody()
method.
MathML support has been added to the HTML parser section. (SVG support is still awaiting input from the SVG WG.)
Web developer-defined attributes have been added. Web developers can add
attributes to elements in the form of
data-name and can access these through the DOM
using dataset[name] on the element in question.
The q element has changed to require punctuation
inside rather than having the browser render it.
The target attribute can now have the value
_blank.
The showModalDialog API has been added.
The document.domain API has been defined.
The source element now has a new
pixelratio attribute useful for videos that have some kind
encoding error.
bufferedBytes, totalBytes and
bufferingThrottled IDL attributes have been added to the
video element.
Media begin event has been renamed to
loadstart for consistency with the Progress Events
specification.
charset attribute has been added to
script.
The iframe element has gained the sandbox
and seamless attributes which provide sandboxing
functionality.
The ruby, rt and rp
elements have been added to support ruby annotation.
A showNotification() method has been added to show
notification messages to the user.
Support for beforeprint and afterprint
events has been added.
The editors would like to thank Ben Millard, Bruce Lawson, Cameron McCormack, Charles McCathieNevile, Dan Connolly, David Håsäther, Dennis German, Frank Ellermann, Frank Palinkas, Futomi Hatano, Gordon P. Hemsley, Henri Sivonen, James Graham, Jens Meiert, Jeremy Keith, Jukka K. Korpela, Jürgen Jeka, Krijn Hoetmer, Leif Halvard Silli, Maciej Stachowiak, Mallory van Achterberg, Marcos Caceres, Mark Pilgrim, Martijn Wargers, Martin Leese, Martyn Haigh, Masataka Yakura, Michael Smith, Mike Taylor, Ms2ger, Olivier Gendrin, Øistein E. Andersen, Philip Jägenstedt, Philip Taylor, Randy Peterman, Steve Faulkner, Toby Inkster, Xaxio Brandish, and Yngve Spjeld Landro for their contributions to this document as well as to all the people who have contributed to HTML over the years for improving the Web!