Home - News ( United States | United Kingdom | Italy | Germany ) - Football scores

Showing content from https://docs.gitlab.com/development/fe_guide/accessibility/best_practices/ below:

Accessibility best practices | GitLab Docs

Since no ARIA is better than bad ARIA, review the following recommendations before using aria-*, role, and tabindex. Use semantic HTML, which has accessibility semantics baked in, and ideally test with relevant combinations of screen readers and browsers.

In WebAIM’s accessibility analysis of the top million home pages, they found that “ARIA correlated to higher detectable errors”. It is likely that misuse of ARIA is a big cause of increased errors, so when in doubt don’t use aria-*, role, and tabindex and stick with semantic HTML.

By default, macOS limits the tab key to Text boxes and lists only. To enable full keyboard navigation:

1. Open System Preferences.
2. Select Keyboard.
3. Open the Shortcuts tab.
4. Enable the setting Use keyboard navigation to move focus between controls.

You can read more about enabling browser-specific keyboard navigation on a11yproject.

• Text, textarea, select, checkbox, radio, file, and toggle inputs have accessible names.
• Buttons, links, and images have descriptive accessible names.
• Icons
  • Non-decorative icons have an aria-label.
  • Clickable icons are buttons, that is, <gl-button icon="close" /> is used and not <gl-icon />.
  • Icon-only buttons have an aria-label.
• Interactive elements can be accessed with the Tab key and have a visible focus state.
• Elements with tooltips are focusable using the Tab key.
• Are any role, tabindex or aria-* attributes unnecessary?
• Can any div or span elements be replaced with a more semantic HTML element like p, button, or time?

Headings are the primary mechanism used by screen reader users to navigate content. Therefore, the structure of headings on a page should make sense, like a good table of contents. We should ensure that:

• There is only one h1 element on the page.
• Heading levels are not skipped.
• Heading levels are nested correctly.

To provide markup with accessible names, ensure every:

• input has an associated label.
• button and link have visible text, or aria-label when there is no visible text, such as for an icon button with no content.
• image has an alt attribute.
• chart has a long and short description.
• fieldset has legend as its first child.
• figure has figcaption as its first child.
• table has caption as its first child.

Remember that an alt attribute should not be longer than approximately 150 characters. While there’s no official guidelines on the length, some screen readers will not read longer strings inside the alt attribute.

An accessible name can be provided in multiple ways and is decided with accessible name calculation. Here is the simplified order of different techniques taking precedence:

1. aria-labelledby
2. aria-label
3. alt, legend, figcaption or caption
4. title.

The following subsections contain examples of markup that render HTML elements with accessible names.

Note that when using GlFormGroup:

• Passing only a label prop renders a fieldset with a legend containing the label value.
• Passing both a label and a label-for prop renders a label that points to the form input with the same label-for ID.

Groups of checkboxes and radio inputs should be grouped together in a fieldset with a legend. legend gives the group of checkboxes and radio inputs a label.

If the label, child text, or child element is not visually desired, use the class name gl-sr-only to hide the element from everything but screen readers.

File input examples:

<!-- File input with a label -->
<label for="attach-file">{{ __('Attach a file') }}</label>
<input id="attach-file" type="file" />

<!-- File input with a hidden label -->
<label for="attach-file" class="gl-sr-only">{{ __('Attach a file') }}</label>
<input id="attach-file" type="file" />

Image examples:

<img :src="imagePath" :alt="__('A description of the image')" />

<!-- SVGs implicitly have a graphics role so if it is semantically an image we should apply `role="img"` -->
<svg role="img" :alt="__('A description of the image')" />

<!-- A decorative image, hidden from screen readers -->
<img :src="imagePath" :alt="" />

Buttons and links should have accessible names that are descriptive enough to be understood in isolation.

<!-- bad -->
<gl-button @click="handleClick">{{ __('Submit') }}</gl-button>

<gl-link :href="url">{{ __('page') }}</gl-link>

<!-- good -->
<gl-button @click="handleClick">{{ __('Submit review') }}</gl-button>

<gl-link :href="url">{{ __("GitLab's accessibility page") }}</gl-link>

In general, avoid using role. Use semantic HTML elements that implicitly have a role instead.

Keyboard users rely on focus outlines to understand where they are on the page. Therefore, if an element is interactive you must ensure:

• It can receive keyboard focus.
• It has a visible focus state.

Use semantic HTML, such as a (GlLink) and button (GlButton), which provides these behaviours by default.

Keep in mind that:

• Tab and Shift-Tab should only move between interactive elements, not static content.
• When you add :hover styles, in most cases you should add :focus styles too so that the styling is applied for both mouse and keyboard users.
• If you remove an interactive element’s outline, make sure you maintain visual focus state in another way such as with box-shadow.

See the Pajamas Keyboard-only page for more detail.

Prefer no tabindex to using tabindex, since:

• Using semantic HTML such as button (GlButton) implicitly provides tabindex="0".
• Tabbing order should match the visual reading order and positive tabindexs interfere with this.

Use interactive elements instead of div and span tags. For example:

• If the element should be clickable, use a button (GlButton).
• If the element should be text editable, use an input or textarea.

Once the markup is semantically complete, use CSS to update it to its desired visual state.

<!-- bad -->
<div role="button" tabindex="0" @click="expand">Expand</div>

<!-- good -->
<gl-button class="gl-p-0!" category="tertiary" @click="expand">Expand</gl-button>

Interactive elements are already tab accessible so adding tabindex is redundant.

<!-- bad -->
<gl-link href="help" tabindex="0">Help</gl-link>
<gl-button tabindex="0">Submit</gl-button>

<!-- good -->
<gl-link href="help">Help</gl-link>
<gl-button>Submit</gl-button>

Screen readers can read text that is not tab accessible. The use of tabindex="0" is unnecessary and can cause problems, as screen reader users then expect to be able to interact with it.

<!-- bad -->
<p tabindex="0" :aria-label="message">{{ message }}</p>

<!-- good -->
<p>{{ message }}</p>

Always avoid using tabindex="1" or greater.

Icons can be split into three different types:

• Icons that are decorative
• Icons that convey meaning
• Icons that are clickable

Icons are decorative when there’s no loss of information to the user when they are removed from the UI.

As the majority of icons within GitLab are decorative, GlIcon automatically hides its rendered icons from screen readers. Therefore, you do not need to add aria-hidden="true" to GlIcon, as this is redundant.

<!-- unnecessary: gl-icon hides icons from screen readers by default -->
<gl-icon name="rocket" aria-hidden="true" />

<!-- good -->
<gl-icon name="rocket" />

Icons convey information if there is loss of information to the user when they are removed from the UI.

An example is a confidential icon that conveys the issue is confidential, and does not have the text “Confidential” next to it.

Icons that convey information must have an accessible name so that the information is conveyed to screen reader users too.

<!-- bad -->
<gl-icon name="eye-slash" />

<!-- good -->
<gl-icon name="eye-slash" :aria-label="__('Confidential issue')" />

Icons that are clickable are semantically buttons, so they should be rendered as buttons, with an accessible name.

<!-- bad -->
<gl-icon name="close" :aria-label="__('Close')" @click="handleClick" />

<!-- good -->
<gl-button icon="close" category="tertiary" :aria-label="__('Close')" @click="handleClick" />

Use the following table to hide elements from users, when appropriate.

To reduce noise for screen reader users, hide decorative images using alt="". If the image is not an img element, such as an inline SVG, you can hide it by adding both role="img" and alt="".

gl-icon components automatically hide their icons from screen readers so aria-hidden="true" is unnecessary when using gl-icon.

<!-- good - decorative images hidden from screen readers -->

<img src="decorative.jpg" alt="">

<svg role="img" alt="" />

<gl-icon name="epic" />

No ARIA is required when using semantic HTML, because it already incorporates accessibility.

However, there are some UI patterns that do not have semantic HTML equivalents. General examples of these are dialogs (modals) and tabs. GitLab-specific examples are assignee and label dropdowns. Building such widgets require ARIA to make them understandable to screen readers. Proper research and testing should be done to ensure compliance with WCAG.

RetroSearch is an open source project built by @garambo | Open a GitHub Issue

Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo

HTML: 3.2 | Encoding: UTF-8 | Version: 0.7.4