This document contains best practice guidance for authors of HTML [[HTML5]] documents on providing text alternatives for images.

This document was developed through the HTML Accessibility Task Force, and is published by the HTML Working Group with approval by the Protocols and Formats Working Group.

It is a draft document and its contents are subject to change without notice.

The body of this document is expected to be incorporated into HTML 5.1 directly.

A list of existing bugs is available. Comments on this draft text should be filed at the W3C Bugzilla using the HTML WG Product and the LC1 alt techniques component, preferably with the a11y keyword.

Introduction

The HTML alt attribute can be placed on any img element. Its content should provide a short textual description of the image. Such descriptions help users who for any reason are unable to view the images. They are also used by software processing HTML documents, such as Web search engines and in-page user agent searches.

Except where otherwise specified the alt attribute MUST be specified on every img element and MUST NOT be empty. The value SHOULD be an appropriate functional replacement for the image.

Providing text to act as an alternative for images

The alt attribute is an essential part of making accessible documents as well as providing for text-based search engines and other consumers of HTML. Text such as the content of alt attributes intended to be used when images are unavailable is known as alternative text, and is an instance of a more general class of Text alternatives [[WCAG20]], a primary way of making visual information accessible. Text alternatives allow the information to be rendered in a variety of ways by a variety of user agents. For example, a person who cannot see a picture can have the text alternative read aloud using synthesized speech.

Authors use text alternatives to provide the same information (or in many cases functionality) to people who cannot see an image. The function of an image depends on its context: a company logo, for example, might be part of identity branding on a home page but function as a navigation icon on other pages. Thus, the same image may need different alt text in different contexts. See also the longdesc attribute [[HTML-LONGDESC]] for a way to link to external text alternatives and image descriptions.

The following sections give some principles for writing text alternatives with the alt attribute and suggest when to use the longdesc attribute to provide more detail. Additional guidance, techniques, and examples are available in Resources on Alternative Text for Images.

Technical Constraints

Because alt is an HTML attribute there are some restrictions on what it can contain and how it can be used:

Language of Alternate Text

The language of the attribute value cannot be marked as being different from that of the image element itself. This is not usually a problem in practice unless the image element was supplied by automated language-based document assembly; use the lang (and for the XML syntax the xml:lang attribute) to mark the language of the alternate text if it differs from that of the surrounding text.

Sub-elements not available in Alternate Text

Attributes are plain text and cannot contain sub-elements. In particular this also means that fragments of MathML or SVG cannot be included in an alt attribute value, and that emphasis cannot be used.

Effective Length Limits on Alternate Text

The value of the alt attribute will generally be rendered by an HTML user agent as replacement text in place of a broken or unavailable image, or if image display is disabled, or in a text-mode browser. The replacement text will usually be rendered in the space reserved for the image. If the text does not fit in the available space, some user agents will truncate the text and others might not display the text at all, reducing accessibility and usability of the document for sighted users in the case that images were not available.

In at least some user agents the CSS overflow property may be used to allow the alternate text to extend outside the width and height allocated for the img element.

Why are we mentioning something that only works in some user agents? I would strike this sentence - we should not be recommending techniquies that don't work portably.

No differentiation from surrounding text

Text readers will generally read out the value of the alt attribute every time it is encountered, and may or may not provide the user with a way to skip to the end of the alternate text and to carry on reading out the document just after the image element. This means that documents should not contain overly burdensome alternate text even if it might fit in the available space.

For visual differentiation, CSS styles such as color can be applied to the img element.

Text that is for use as an alternative to an image but that contains formatting or markup (e.g., tables or equations) needs to be made available separately as described in the next section, Principles for providing text alternatives. In this case a short text alternative (e.g., Maxwell's Equations) SHOULD be used in the alt attribute.

The alt attribute may be subject to XML “attribute value normalization” if the document is served as XML or processed by XML tools. In particular this normalization removes leading and trailing ASCII whitespace. Non-breaking spaces such as   are not subject to normalization and can be used when concatenating the alt text value to surrounding text might cause unacceptable word joining.

Principles For Providing Text Alternatives

Even though there are restrictions on its value, the alt attribute is still needed in order to make documents properly accessible. This section describes general principles for authors and developers to use when considering the use of the alt attribute, the longdesc attribute, and also the newer ARIA described-by and described-at attributes. [[WAI-ARIA]]

We mention described-at here, but it is 1) not approved and 2) we don't point that out until much later in the document. We should probably remove this reference or at least point to the information about its status.

The Web Accessibility Initiative provides An fDecision Tree to help authors decide what to put into an alt attribute value.

Functional images: State the Purpose

Where an image has a function (e.g., a button to press) or serves a clear purpose such as identifying the maker of a page by a company logo, be direct and simply give the purpose (e.g., Make Payment or W3C Logo).

Images As Text: Give the text

If an image is used to stand for a special character, emoji, gaiji, itaiji, emoticon, or text in an unusual font, the alt text should simply be the text that is on the image, or a textual equivalent such as :-) for a smiling face.

In many cases Web fonts can be used instead of images, and, provided that the fonts are correctly encoded and that the characters represented are in Unicode, user agents that do not use the fonts will still render the correct text.

Images of Equations and Tables

Recommended practice is to mark up mathematical equations with MathML. If this is not possible, sometimes a picture of an equation is used. Similarly even complex tables can usually be marked up in HTML directly, but sometimes a picture is used instead. In these and similar cases the alt attribute should be given a value such as Equation 7 or Table of wandering monsters by level. Since people seeing the image would not want to read the same content in two different forms, the information in the equation or table picture should then be presented in an external document, pointed to with a URI as the value of the longdesc attribute, for example longdesc = "equation007.html".

Charts and Diagrams

It is often possible to represent charts, diagrams, and other explanatory sorts of pictures using [[SVG]]. This does not automatically make such pictures accessible, but different mechanisms then become available for embedding text alternatives in the SVG images.

If SVG is for some reason not used then the same considerations as for images of equations and tables generally apply. A description of a diagram, however, may need to include information about the presentation as well as the information. For example, a government report in an archival format might mention the colors used in a bar chart in order to facilitate subsequent discussions.

Simple Images

Where an image or diagram can be described adequately with a short piece of plain text that would easily fit in the space otherwise occupied by the image, that text should be used for the text alternative.

Decorative and Pieced Images

If an image is purely decorative, it is not necessary to give it alt text; use an empty value for the alt attribute. If the image separates words of text an alternative text of an encoded single space (e.g., alt="") might be appropriate.

When multiple images are placed together, for example to make up a frame or border or when a complex image is split into smaller images, only the first image should normally have explanatory text, in order to avoid tedious repetition from a screen reader.

Examples of scenarios where users benefit from text alternatives for images

Common text alternative techniques

When an a element that is a hyperlink, or a button element, has no text content but contains one or more images, include text in the alt attribute(s) that together convey the purpose of the link or button.

5 buttons: bold, italic, strike through, bulleted list and numbered list.

  <ul>
  <li><button><img src="b.png" alt="Bold"></button></li>
  <li><button><img src="i.png" alt="Italics"></button></li>
  <li><button><img src="strike.png" alt="Strike through"></button></li>
  <li><button><img src="blist.png" alt="Bulleted list"></button></li>
  <li><button><img src="nlist.png" alt="Numbered list"></button></li>
  </ul>
  

The image link is broken, and I don't see such an image name in the current HTML source. Where is this image?

In this example, a link contains a logo. The link points to the W3C web site from an external site. The text alternative is a brief description of the link target.

W3C logo used as link content.

<a href="http://w3.org">
  <img src="images/w3c_home.png" width="72" height="48" alt="W3C web site">
</a>    

This example is the same as the previous example, except that the link is on the W3C web site. The text alternative is a brief description of the link target.

W3C logo used as link content.

<a href="http://w3.org">
  <img src="images/w3c_home.png" width="72" height="48" alt="W3C home">
</a>  

Depending on the context in which an image of a logo is used it could be appropriate to provide an indication, as part of the text alternative, that the image is a logo. Refer to section on Logos, insignia, flags, or emblems.

In this example, a link contains a print preview icon. The link points to a version of the page with a print stylesheet applied. The text alternative is a brief description of the link target.

Print preview icon used as link content.

<a href="preview.html">
  <img src="images/preview.png" width="32" height="30" alt="Print preview.">
  </a>  

In this example, a company logo for the PIP Corporation has been split into the following two images, the first containing the word PIP and the second with the abbreviated word CO. The images are the sole content of a link to the PIPCO home page. In this case a brief description of the link target is provided. As the images are presented to the user as a single entity the text alternative PIP CO home is in the alt attribute of the first image.

Image containing the text 'PIP'.Image containing the text 'CO'.

<a href="pipco-home.html">
  <img src="pip.gif" alt="PIP CO home"><img src="co.gif" alt="">
  </a>    

Graphical Representations: Charts, diagrams, graphs, maps, illustrations

Readers can benefit when content is presented in graphical form, for example as a flowchart, a diagram, a graph, or a map showing directions. People unable to see the image for any reason will need the information provided in another form.

Diagrams often convey a lot of information. This means that a textual alternative may be quite lengthy. The content of the alt attribute is in general limited to the amount of text that will fit in the space normally occupied by the image. In addition, it may be tedious for someone listening to an audio rendition of a Web page to have to skip over a lengthy description each time they reach the diagram.

An appropriate technique, then, is to have a short description in the alt attribute, such as bar chart of product sales by region, and to link to a longer image description.

An image description in this context is text, whether embedded in the main document or elsewhere, that would enable a reader to have a good idea of what the image looks like. This is different from an alternate representation in that, for example, it might mention that a pie chart is circular or that the items in a bar chart are horizontal rows, and should mention any significant colors. Such a description can enable people to participate in conversations about the image, or to answer examination questions that might rely on the way the information is represented and not just the underlying data.

Using the alt attribute with an external image description

An external image description is just another document, normally in HTML, that describes an image. You should consider using an external image description if one of more of the following applies:

Using the alt attribute with an associated text alternative

A text alternative presents the same essential information as a diagram or illustration but in a text format accessible in cases where the image is not seen or used. It is normally contained in the same document as the image. You should supply a text alternative in addition to the alt attribute in these circumstances:

The following example is taken from a (fictional) examination question on simple color theory. Someone unable to see the image would need to know that the two circles are colored red and blue respectively, and that the white area is where the circles overlap. Someone able to see the image would not want to waste time in an exam context reading such a description. In addition, the description is too long for an alt attribute value.

To meet the needs both of users who see the image an users who don't, an image description is provided separately. In some contexts this could be included in the same HTML document as the examination question but image descriptions for educational purposes are often provided by third parties and may be more easily managed as a separate file. In addition browsers used by people with accessibility needs might not have JavaScript available or enabled. In this example the widely-supported longdesc attribute is used to point to an external description; a more general attribute to point to external descriptions, aria-describedat, has been proposed but is not yet (2015) widely supported.

Is there really a reason to say whether longdesc is widely supported or not? Longdesc is part of HTML5 so we can safely recommend its use. We don't need to point out that it is widely supported. It has been in HTML for ~20 years - Shane

overlapping circles
What color should be used to complete the white part of the diagram, assuming a reflective color model?

<p><img src="overlapping-circles.html"
	  alt="overlapping circles" longdesc="overlapping circles-d.html><br>
	What color should be used to complete the white part of
	the diagram, assuming a reflective color model?</p>
	  

Using the alt attribute with an associated text alternative

You should consider supplying a text alternative in addition to the alt attribute in these circumstances:

Although the alt attribute is essential to enable someone to understand which image was not rendered, it should not be relied on to convey detailed content.

Here's an example of a flowchart image with a short text alternative included in the alt attribute. In this case the text alternative is a description of the link target as the image is the sole content of a link. The link points to a description (within the same document) of the process represented in the flowchart.

Flowchart: Dealing with a broken lamp.

<a href="#desc"><img src="flowchart.gif" alt="Flowchart: Dealing with a broken lamp."></a>
...
...
<div id="desc">
<h2>Dealing with a broken lamp</h2>    
<ol>    
<li>Check if it's plugged in, if not, plug it in.</li>
<li>If it still doesn't work; check if the bulb is burned out. If it is, replace the bulb.</li>
<li>If it still doesn't work; buy a new lamp.</li>    
</ol>
</div>

In this example, there is an image of a chart. It would be inappropriate to provide the information depicted in the chart as a plain text alternative in an alt attribute as the information is a data set. Instead a structured text alternative is provided below the image in the form of a data table using the data that is represented in the chart image.

Bar Chart showing average rainfall in millimetres by Country and Season.

Indications of the highest and lowest rainfall for each season have been included in the table, so trends easily identified in the chart are also available in the data table.

Average rainfall in millimeters by country and season.
United Kingdom Japan Australia
Spring 5.3 (highest) 2.4 2 (lowest)
Summer 4.5 (highest) 3.4 2 (lowest)
Autumn 3.5 (highest) 1.8 1.5 (lowest)
Winter 1.5 (highest) 1.2 1 (lowest)
<figure role="group" aria-labelledby="caption">
<img src="rainchart.gif" alt="Bar chart: Average rainfall in millimeters by Country and Season.">
<table>
<caption id="caption">Rainfall in millimeters by Country and Season.</caption>
<tr><td><th scope="col">UK <th scope="col">Japan<th scope="col">Australia</tr>
<tr><th scope="row">Spring <td>5.5 (highest)<td>2.4 <td>2 (lowest)</tr>
<tr><th scope="row">Summer <td>4.5 (highest)<td>3.4<td>2 (lowest)</tr>
<tr><th scope="row">Autumn <td>3.5 (highest) <td>1.8 <td>1.5 (lowest)</tr>
<tr><th scope="row">Winter <td>1.5 (highest) <td>1.2 <td>1 lowest</tr>
</table>
</figure>    

The table image and data table are grouped using a figure element with a [[WAI-ARIA]] role="group". The figure is labelled by the table caption content using aria-labelledby.

Once figure and figcaption are accessibility supported in browsers, the addition of ARIA attributes will not be required.

For any of the examples in this section the details and summary elements could be used so that the text descriptions for the images are only displayed on demand:

Default (closed) view:

Details element in the closed state.

Expanded view (after control activated):

Details element in the open state with list content displayed.

  <figure>
  <img src="flowchart.gif" alt="Flowchart: Dealing with a broken lamp.">
  <details>
  <summary>Dealing with a broken lamp</summary>
  <ol>
  <li>Check if it's plugged in, if not, plug it in.</li>
  <li>If it still doesn't work; check if the bulb is burned out. If it is, replace the bulb.</li>
  <li>If it still doesn't work; buy a new lamp.</li>
  </ol>
  </details>
  </figure>
  

The details and summary elements are not well supported by browsers. Until they are well supported, you will need to use scripting to provide the functionality. There are a number of scripted polyfills and custom controls available in popular JavaScript UI widget libraries that provide similar functionality.

Images of text

Sometimes an image contains only text, and the purpose of the image is to display text using visual effects and/or fonts. It is strongly recommended that text styled using CSS be used, but if this is not possible, provide the same text in the alt attribute as is in the image.

This example shows an image of the text "Get Happy!" written in a fancy multi colored freehand style. The image makes up the content of a heading. In this case the text alternative for the image is "Get Happy!".

Get Happy!

<h1><img src="gethappy.gif" alt="Get Happy!"></h1>

When an image is used to represent a character that cannot otherwise be represented in Unicode (e.g., gaiji, itaiji, or new characters such as novel currency symbols) the text alternative should be a more conventional way of writing the same thing (e.g., using the phonetic hiragana or katakana to give the character's pronunciation).

An image should not be used if Unicode characters would serve an identical purpose. Only when the text cannot be directly represented using Unicode (e.g., because of decorations or because the character is not in the Unicode character set) would an image be appropriate. If necessary Web fonts can be used to supply specific representations for Unicode characters. This is preferable to the use of images because a text reader or search engine can process the Unicode character directly.

An illuminated manuscript might use ornate forms for some of its letters. The text alternative in such a situation is just the character that the image represents.

Once upon a time and a long long time ago...

<p><img src="initials/fancyO.png" alt="O">nce upon a time and a long long time ago...

A graphical representation of some of the surrounding text

In many cases an image is actually just supplementary &emdash; its presence merely reinforces the surrounding text. In these cases, the alt attribute must be present but its value MUST be the empty string.

In general, an image falls into this category if removing the image doesn't make the page any less useful, but including the image makes it a lot easier for users of visual browsers to understand the concept.

It is not always easy to write a useful text alternative for an image. Another option is to provide a link to a description or further information about the image when one is available.

In this example of the same image, there is a short text alternative included in the alt attribute, and there is a link after the image. The link points to a page containing information about the painting.

The Lady of Shalott
A poem by Alfred Lord Tennyson.

Painting of a woman in a white flowing dress, sitting in a small boat.

About this painting.

Full recitation of Alfred, Lord Tennyson's poem.

<header>
<h1>The Lady of Shalott</h1>
<p>A poem by Alfred Lord Tennyson</p>
</header>
<figure>
<img src="shalott.jpeg" alt="Painting of a woman in a white flowing dress, sitting in a small boat.">
<p><a href="http://bit.ly/5HJvVZ">About this painting.</a></p>
</figure>

<!-- Full Recitation of Alfred, Lord Tennyson's Poem.  -->

A purely decorative image that doesn't add any information

Purely decorative images are visual enhancements, decorations, or embellishments that provide no function or information beyond aesthetics to users who can view the images.

Mark purely decorative images so they can be ignored by assistive technology by using an empty alt attribute (alt="").

Here's an example of an image being used as a decorative banner for a person's blog. The image offers no information and so an empty alt attribute is used.

Clara's Blog

Welcome to my blog...

<header>
<div><img src="border.gif" alt="" width="400" height="30"></div>
<h1>Clara's Blog</h1>
</header>

<p>Welcome to my blog...</p>  

Inline images

When images are used inline as part of the flow of text in a sentence, provide a word or phrase as a text alternative which makes sense in the context of the containing sentence:

I love you.

I <img src="heart.png" alt="love"> you.

My heart breaks.

My <img src="heart.png" alt="heart"> breaks.

Unicode characters could be used as an alternative to inline images where appropriate. Modern browsers have support for characters like "♥" and "☺", and support the use of custom fonts to render those as colored image fonts.

Image maps

If an img element has a usemap attribute which references a map element containing area elements that have href attributes, the img is considered to be interactive content. In such cases, always provide a text alternative for the image using the alt attribute.

Consider the following image which is a map of Katoomba, it has 2 interactive regions corresponding to the areas of North and South Katoomba:

Map of Katoomba.

North Katoomba South Katoomba

The text alternative is a brief description of the image. The alt attribute on each of the area elements provides text describing the content of the target page of each linked region:

<p>View houses for sale in North Katoomba or South Katoomba:</p>
<p><img src="imagemap.png" width="209" alt="Map of Katoomba" height="249" usemap="#Map">

<map name="Map">
<area shape="poly" coords="78,124,124,10,189,29,173,93,168,132,136,151,110,130" href="north.html"
alt="Houses in North Katoomba">
<area shape="poly" coords="66,63,80,135,106,138,137,154,167,137,175,133,144,240,49,223,17,137,17,61"
alt="Houses in South Katoomba" href="south.html">
</map> 

Generally, slicing an image for links should be used instead of image maps as fallback support for image maps is poorly supported.

When a text alternative is not available at the time of publication

In some cases an image is included in a published document, but the author is unable to provide an appropriate text alternative. In such cases the minimum requirement is to provide a caption for the image using the figure and figcaption elements under the following conditions:

  • The img element is in a figure element.
  • The figure element contains a figcaption element.
  • The figcaption element contains content other than inter-element whitespace.
  • Ignoring the figcaption element and its descendants, the figure element has no Text node descendants other than inter-element whitespace, and no embedded content descendant other than the img element.

In other words, the only content of the figure is an img element and a figcaption element, and the figcaption element must include (caption) content.

Such cases SHOULD be kept to an absolute minimum. If there is even the slightest possibility of the author having the ability to provide real alternative text, then it would not be acceptable to omit the alt attribute.

In this example, a person uploads a photo, as part of a bulk upload of many images, to a photo sharing site. The user has not provided a text alternative or a caption for the image. The site's authoring tool inserts a caption automatically using whatever useful information it has for the image. In this case it's the file name and date the photo was taken.

The caption text in the example below is not a suitable text alternative and is not conforming to the Web Accessibility Guidelines 2.0. [[WCAG20]]

no <a>text alternative</a> provided

clara.jpg, taken on 12/11/2010.

<figure>
<img src="clara.jpg">
<figcaption>clara.jpg, taken on 12/11/2010.</figcaption>
</figure>    

Notice that even in this example, as much useful information as possible is still included in the figcaption element.

In this second example, a person uploads a photo to a photo sharing site. She has provided a caption for the image but not a text alternative. This may be because the site does not provide users with the ability to add a text alternative in the alt attribute.

no <a>text alternative</a> provided

Eloisa with Princess Belle

<figure>
<img src="elo.jpg">
<figcaption>Eloisa with Princess Belle</figcaption>
</figure>    

CAPTCHA Images

CAPTCHA stands for "Completely Automated Public Turing test to tell Computers and Humans Apart". CAPTCHA images are used for security purposes to confirm that content is being accessed by a person rather than a computer. This authentication is done through visual verification of an image. CAPTCHA typically presents an image with characters or words in it that the user is to re-type. The image is usually distorted and has some noise applied to it to make the characters difficult to read.

To improve the accessibility of CAPTCHA provide text alternatives that identify and describe the purpose of the image, and provide alternative forms of the CAPTCHA using output modes for different types of sensory perception. For instance provide an audio alternative along with the visual image. Place the audio option right next to the visual one. This helps but is still problematic for people without sound cards, the deaf-blind, and some people with limited hearing. Another method is to include a form that asks a question along with the visual image. This helps but can be problematic for people with cognitive impairments.

It is strongly recommended that alternatives to CAPTCHA be used, as all forms of CAPTCHA introduce unacceptable barriers to entry for users with disabilities. For further information, see [[TURINGTEST]].

This example shows a CAPTCHA test which uses a distorted image of text. The text alternative in the alt attribute provides instructions for a user in the case where she cannot access the image content.

captcha containing the words 'aides' and 'sprucest'. The letters are distorted and the color of the letters and background is partially inverted,

<img src="captcha.png" alt="If you cannot view this image an audio challenge is provided.">
 <!-- audio CAPTCHA option that allows the user to listen and type the word -->
 <!-- form that asks a question --> 

Methods for Providing Text Alternatives

Native HTML methods

The img element alt attribute

The primary method for providing text alternatives for images is by including text in the img element alt attribute. In graphical browsers the content of the alt attribute is typically displayed along with an indication (bordered area and/or an image icon) of the presence of an image when the image is not displayed, because the user has disabled image display or the image source information is incorrect. Assistive technologies such as screen readers will typically announce the presence of an image followed by the alt attribute content. Text based browsers may display the alt attribute content in brackets or in different colored text or as the content of a link to the image or as text without any indication of the image or prefixed with an indication of the image.

<img src="example.jpg" alt="text alternative.">
Advantages:
  • Robust support in most graphical & non graphical web browsers and assistive technologies.
  • Text alternative is explicitly associated with the image.
  • Inclusion of text alternatives as alt attribute content does not impact upon the default visual display of content.
Disadvantages:
  • Content of the alt attribute is not available to some users who may find it useful.
  • If the alt attribute content is too long it may cause display issues in some browsers.
  • If the alt attribute content is too long it may cause reading issues in some assistive technologies.
  • The entire contents of the alt attribute are read by assistive technologies every time the image is encountered, with no possibility of skipping and no distinction from the surrounding text.
  • Semantic structure cannot be added to alt attribute content.

Is alt attribute content a replacement for an image?

The answer to this question depends on the content of the image and the context the image is being used in:

  • When an image contains too much information to use the alt attribute as a container for a text alternative, the alt attribute content can be a label for the image. This label MAY be a brief description that identifies the image. This SHOULD be accompanied by a more complete text alternative that is programmatically associated with the image.
  • When an image contains structured information that cannot be conveyed using the alt attribute as a container for a text alternative, the alt attribute content SHOULD be a label for the image. This label MAY be a brief description that identifies the image. This SHOULD be accompanied by a structured text alternative that is programmatically associated with the image.
  • When an image is used to represent text and it is the author's intent that the purpose of using an image of text is to achieve a visual style, the alt attribute content SHOULD be considered as a replacement for the image.
  • When an image is decorative and it is the author's intent that it not convey any information, an empty alt attribute SHOULD be considered as an indication that the image can be safely ignored.
  • When it is the author's intent that the image is not to be seen by users, an empty alt attribute SHOULD be considered as an indication that the image can be safely ignored.
  • When an image is immediately preceded or followed by a text alternative and the image and the text alternative are not complex in nature, an empty alt attribute SHOULD be considered as an indication that the image can be safely ignored.
  • When an image is the sole content of a link, the alt attribute content SHOULD be a brief description of the link target.
  • When an image is the content of a link and is immediately preceded or followed by a brief description of the link target, the alt attribute content SHOULD be empty.
  • When an image is the content of a link that also contains structured text content, if the text content is a description of the link target, the alt attribute content MAY be empty or a text alternative for the image. If the text content is not a description of the link target, the alt attribute content SHOULD be a brief description of the link target.

Using an empty alt attribute alt=""

Circumstances in which it is appropriate to use an empty or null alt attribute:

Circumstances in which it is not appropriate to use an empty or null alt attribute:

  • An image is contained within a figure element and has an associated caption provided using the figcaption element. Using an empty alt attribute hides an image from some users. If an image has a caption the image needs to be discoverable by users, otherwise a caption is present that refers to nothing for some users.
  • An image contains relevant information not available in the same document as text. Using an empty alt attribute hides an image from some users. In order to have the possibility of interrogating the image, the image must be discoverable.
  • An image contains relevant information, an alternative interpretation of which is available in the same document as structured text. Using an empty alt attribute hides an image from some users, which is incorrect. The image is not meaningless, it contains information which a range of users could interpret with the aid of the short text alternative and longer description. It also provides a text alternative for users who have images turned off in their browsers, so they can load and view the image if they wish. If an empty alt attribute is present there may be no indication that an image is present. Furthermore if a description of an image is provided in a document, a programmatic association between the image and the descriptive text is required. Using an empty alt attribute on the image effectively precludes the assigning of a programmatic association.

How long should a text alternative be?

A text alternative for an image should be as long as it needs to be to adequately convey the information in the image in the context in which the image is being used.

How long should an alt text be?

While there are no definitive right or wrong lengths for text alternatives provided using the img element's alt attribute, the general consensus is that if the text alternative is longer 30-50 words (2 to 3 sentences), it should not be considered a short text alternative and should not be presented using the alt attribute. Refer to the section - Graphical Representations: Charts, diagrams, graphs, maps, illustrations for example methods of providing longer text alternatives for images.

A text alternative provided using the alt attribute is exposed to screen reader users as a text string, generally announced as a chunk, and cannot be structured into paragraphs or have other markup added to aid comprehension. It also cannot usually be interrupted, so brevity can be of considerable value.

The figure and figcaption elements

The figure and figcaption elements provide a method to explicitly associate a caption with a variety of content including images. Any content inside the figure element that is not contained within the figcaption element is labelled by the content of the figcaption element. The figcaption content may be an adjunct to the text alternative provided using the alt attribute:

The figcaption content may be a text alternative for the image, obviating the need for a text alternative provided using the alt attribute. This would only be the case if the figcaption content provides an adequate text alternative for the visual content in the image:

Shadow like figures and a graffiti tag drawn on the walls of a partially demolished building, illuminated by the light from a street lamp.

<img src="example.jpg" alt="Shadow like figures and a graffiti tag drawn on the walls of a partially
demolished building, illuminated by the light from a street lamp.">

(photo)

Drawings on the walls of a partially demolished building (photo)

<figure>   
<img src="shadows.jpg">
<figcaption>Drawings on the walls of a partially demolished building (photo). </figcaption>
</figure>
Advantages:
  • When supported by browsers and assistive technology the content of the figcaption will be explicitly associated with the image.
Disadvantages:
  • The figure and figcaption elements are not currently accessibility supported by all browsers.

Until the figure and figcaption elements are widely accessibility supported by browsers and/or assistive technology it is recommended that role="group" be used on the figure element and in the case where the figcaption content is the text alternative, provide a label for the image in the alt attribute. These recommendations are illustrated in following example:

(photo 1)

Shadow like figures and a graffiti tag drawn on the walls of a
partially demolished building, illuminated by the light from a street lamp (photo 1).

<figure role="group">
      <img src="shadows.jpg" alt="photo 1">
      <figcaption>Shadow like figures and a graffiti tag drawn on the walls of a
     partially demolished building, illuminated by the light from a street lamp (photo 1).
     </figcaption>
     </figure>

Using nested figure elements

In cases where there is a group of related images each having an associated caption, it is recommended that nested figure and figcaption elements be used.

The castle through the ages: 1423, 1756, and 1966 respectively.

The castle has one tower, and a tall wall around it.

Charcoal on wood. Anonymous, circa 1423

.The castle now has two towers and two walls.

Oil-based paint on canvas. Eloisa Faulkner, 1756.

The castle lies in ruins, the original tower all that remains in one piece.

Film photograph. Séraphin-Médéric Mieusement, 1936.

<figure>
<figcaption>The castle through the ages: 1423, 1756, and 1966 respectively.</figcaption>
<figure>
<img src="castle-etching.jpg" alt="The castle has one tower, and a tall wall around it.">
<figcaption>Charcoal on  wood. Anonymous, circa 1423.</figcaption>
</figure>
<figure>
<img src="castle-painting.jpg" alt="The castle now has two towers and two walls.">
<figcaption>Oil-based paint on canvas. Eloisa Faulkner, 1756.</figcaption>
</figure>
<figure>
<img src="castle-fluro.jpg"
alt="The castle lies in ruins, the original tower all that remains in one piece.">
<figcaption>Film photograph. <span lang="fr">Séraphin Médéric Mieusement</span>, 1936.</figcaption>
</figure>
</figure>

The img element title attribute

The title attribute MUST NOT be used to provide a text alternative for an image. The title attribute MUST NOT be used to provide a caption for an image; use the figure and figcaption elements to provide a caption, as described above.

Notes on use of Gaiji

Gaiji is a term meaning "external characters". In general, this means characters that are not represented in existing Japanese encoding systems such as [[UNICODE]].

  • When gaiji is used as a "hack" to display a character that is adequately represented by a real Unicode character, the recommended practice is to remove the gaiji image and just use the Unicode character.
  • When gaiji is used to display a character that is not available in a particular font, the alternative text should be the Unicode character.
  • When gaiji is used to represent a graphical image similar to those displaying emoji, use a short description of the image's content (e.g. "thumbs up", "sunrise", or "smiling woman").
  • When gaiji is used to represent a new or made-up character that has no Unicode representation, use the phonetic hiragana or katakana string to represent the pronunciation of the new or made-up character.

Glossary

Accessibility API

Operating systems and other platforms provide a set of interfaces using an accessibility API that expose information about objects and events to assistive technologies. Assistive technologies use these interfaces to get information about and interact with those widgets. Examples of this are the Java Accessibility API [JAPI], Microsoft Active Accessibility [MSAA], the Mac OS X Accessibility Protocol [AXAPI], the Gnome Accessibility Toolkit (ATK) [ATK], and IAccessible2 [IA2].

Accessible Name

The accessible name is the name of a user interface element. Each platform accessibility API provides the accessible name property. The value of the accessible name may be derived from a visible (e.g., the visible text on a button) or invisible (e.g., the text alternative that describes an icon) property of the user interface element.

alt text

Alt text is a colloquial term for a text alternative provided using the alt attribute of an img element.

Assistive Technology

An assistive technology is a hardware and/or software that acts as a user agent, or along with a mainstream user agent, to provide functionality to meet the requirements of users with disabilities that go beyond those offered by mainstream user agents. A more detailed explanation of assistive technology is provided in the WCAG 2.0 glossary. [[WCAG20]]

Content (Web Content)

Information and sensory experience to be communicated to the user by means of a user agent such as a web browser, including code or markup that defines the content's structure, presentation, and interactions.

Empty alt Attribute or Null alt Attribute

An alt attribute with no content:

<img src="null.gif" alt="">

Non-text Content

A non-text content is any content that is not a sequence of characters that can be programmatically determined or where the sequence is not expressing something in human language

Note: This includes ASCII Art (which is a pattern of characters), emoticons, leetspeak (which uses character substitution), and images representing text.

Programmatic Association

It must be possible for people using assistive technologies to find the text alternative for an image when they encounter the image that they cannot use. To accomplish this, the text must be "programmatically associated" with the image and is referenced as programmatic association. This means that the user must be able to use their assistive technology to find the alternative text (that they can use) when they land on the image (that they can't use).

Text Alternative

Text that is programmatically associated with non-text content or referred to from text that is programmatically associated with non-text content. Programmatically associated text is text whose location can be programmatically determined from the non-text content. [[UNDERSTANDING-WCAG20]]

Acknowledgements

Incomplete and in no particular order - Heydon Pickering, David MacDonald, Laura Carlson, Josh O Connor, Gez Lemon, Anne van Kesteren, Ian Hickson, Mike Smith, Mike Paciello, Bim Egan, Gregory Rosmaita, Michael Cooper, Janina Sajka, Matt May, Bevi Chagnon, Jonathan Avila, Pat Rees, Charlie Pike, Andy Maseyk, Rich Clark.