The trouble with table-based layouts

HTML tables aren’t a good choice to use to create columns or layouts in a webpage because they aren’t as responsive to small screens and can cause confusion to screen reader users. Back in the late 1990s a lot of designers used tables for visual layout, but this was later found to be a bad practice because fixed-width table-based layouts didn’t adaptive well to different-sized screens. This layout practice used the HTML <table> element for a use that it wasn’t semantically intended. HTML tables were designed for data grids, from the perspective of semantic structure, so it is better to use them only for that purpose.

How developers can ensure layout tables don’t cause problems for assistive technology

Web developers should take steps to ensure layout tables don’t cause problems for screen reader users. If an HTML table must be used for a layout, the developer should add the role=”presentation” attribute to the <table> element so it’s not confused with a data table. Role=”presentation” tells the screen reader to ignore the semantic table markup and treat it as normal text. The web developer also should not include any of the header markup for data tables in a purely visual layout table. These data table header markup tags to avoid are:

  • The <caption> element
  • The <summary> attribute
  • The <th> element
  • The scope attribute (that goes in a <th> tag)
  • The headers attribute (that goes in a <td> table cell tag)

If data table markup tags are used in a purely visual layout table, screen reader users might expect the table to represent tabular data leading to confusion or a poor user experience. So, if a purely visual layout table must be used the developer should ensure that ARIA is used to make it look like normal text to screen readers. Any purely decorative images used in the table so be marked with null alt text (alt=” “) or hidden from readers with aria-hidden=”true”. Back in the late 1990s, web designers used to use tools like Macromedia Fireworks to slice up webpage visual mockups into layout tables and many time these tables would include a bunch of purely decorative images.

An easy way to avoid the troubles with layout tables is to avoid using them at all. For the modern web CSS (cascading style sheets) is a more robust way to design visual layouts. Once web developers realized web sites would be viewed on devices of many different screen sizes, they began using CSS to create columns, float content, and adapt to different viewport widths. The latest Web Accessibility Guidelines (WCAG 2.1) say that websites should be responsive and including layout tables would likely work against that. According to WCAG guidelines, tables should reflow to fill most of the width of the viewport, without causing any horizontal viewport overflow. On smaller screens it becomes more difficult avoid horizontal scrolling (horizontal overflow).  The WCAG SC that is related to tables and horizontal viewport overflow is 1.4.10 Reflow. This success criterion requires that content can be presented without loss of information or functionality, and without requiring scrolling in two dimensions for vertical scrolling content at a width of 320 CSS pixels.

How content authors can avoid using tables for layout

As a content author, might craft your blog posts in MS Word documents and you might desire to create a visually interesting layouts using tables in Word. But doing so will introduce some of the same accessibility problems as with using HTML tables for layout in a webpage.  It’s best to paste your content into WordPress blog posts as plain text without MS Word formatting so the conversion to HTML will be more semantically correct, free from accessibility problems. Pasting a table-based visual layout from a Word doc should be avoided because it can introduce accessibility problems into a blog post and lead to poor user experiences for both visual and screen reader users.  Tables aren’t really a valid layout technique in Word documents. They are meant for tabular date (data grids). When publicly shared on the web at public universities, Word docs (and PDFs) are expected to follow WCAG guidelines. So, it’s best to follow accessibility best practices even in our Word documents.

WordPress blog sites may or may not include a table tool in the rich text editor for posts. If your WordPress blog does include a table tool, be sure to use it responsibly. Avoid using the table tool for layout tables unless you have the HTML expertise to markup the table so it doesn’t cause problems for assistive technology. If your blog includes a table tool, it’s best to stick to using it just for data tables. You can ask your department’s web developer to check your data table for accessible markup. It’s possible for rich text editors to create simple accessible data tables but when you get into more complex data tables you’ll have to go to the code view to insert the correct accessible markup. Some more advance table markup that might not be in your rich text editor’s table tool may include:

  • Ability to add the scope attribute (that goes in a <th> tag)
  • Ability to add the headers attribute (that goes in a <td> table cell tag)

Once you get into creating more complex data tables you should probably work with your department’s web developer to make them accessible.  For accessible presentation in websites, sometimes it’s better to split a complex table into two or more simple tables. This makes the data tables easier to understand to screen reader users a more adaptable to different screen sizes (more responsive).

Opening a link in a new window or tab

I tend to like opening links to external websites in a new window with target=_blank. Below is an example of the code I use to make a link like that more accessible. WCAG guidelines say we should warn a screen reader user when a link is going to open in a new window, so I do that with an aria-label attribute in the a tag. Using an aria-label attribute is a way to override the standard accessible name of the link and provide more context for screen reader users.

Example of a link with an aria-label attribute in the a tag

Videos can be hosted in MediaSite

Videos can be hosted in MediaSite

Actually, the best practice for links that open in a new window is to provide both visual and verbal indication. For sighted users an external link icon appended at the end of the link could be used. See this article about opening links in new windows article from the University of California Berkeley for more information about best practices. WCAG guidelines only recommend opening links in new windows in a few special situations.

Screen readers and languages test page

The following paragraphs are marked up with different languages in the lang attribute. These can be used for testing the accent of a screen reader. The two letter language code in the lang attribute should match the language the text was written in.

If a language isn’t specified, the screen reader will read the page in the user’s default language. That may result in a bad accent that’s difficult to understand. To specify a language for the page, add the ‘lang’ attribute to the HTML tag. For example: html lang=en.

Si aucune langue n’est spécifiée, le lecteur d’écran lira la page dans la langue par défaut de l’utilisateur. Cela peut entraîner un mauvais accent difficile à comprendre. Pour spécifier une langue pour la page, ajoutez l’attribut ‘lang’ à la balise HTML. Par exemple : html lang=fr.

Marked up as French lang=”fr” a language isn’t specified, the screen reader will read the page in the user’s default language. That may result in a bad accent that’s difficult to understand. To specify a language for the page, add the ‘lang’ attribute to the HTML tag. For example: html lang=fr.

Wenn keine Sprache angegeben ist, liest der Bildschirmleser die Seite in der Standardsprache des Benutzers vor. Das kann zu einem schlechten Akzent führen, der schwer zu verstehen ist. Um eine Sprache für die Seite anzugeben, fügen Sie dem HTML-Tag das Attribut „lang“ hinzu. Beispiel: html lang=de.

Marked up as German lang=”de”: If a language isn’t specified, the screen reader will read the page in the user’s default language. That may result in a bad accent that’s difficult to understand. To specify a language for the page, add the ‘lang’ attribute to the HTML tag. For example: html lang=de.

>यदि कोई भाषा निर्दिष्ट नहीं है, तो स्क्रीन रीडर पृष्ठ को उपयोगकर्ता की डिफ़ॉल्ट भाषा में पढ़ेगा। इसका परिणाम खराब लहजे में हो सकता है जिसे समझना मुश्किल है। पृष्ठ के लिए एक भाषा निर्दिष्ट करने के लिए, HTML टैग में ‘लैंग’ विशेषता जोड़ें। उदाहरण के लिए: html lang=hi.


ISO 2 Letter Language Codes reference

Google Translate

ISO 639.2 Codes for the Representation of Names of Languages

Complex images and alt text descriptions

Complex images like charts & graphs must be briefly described using basic alt text and must have a more complete long description elsewhere in the page. The long description should be available to both sighted and screen-reader users and be programmatically associated with the image.

The figure / figcaption method:

One of the limitations of basic alt text is that it’s meant to be brief, no more than around 150 characters. So you shouldn’t try to use an alt attribute in an image for the long description. Instead, a good way to add a programmatically associated long description is the use the figure / figcaption method. In this method the figure tag wraps the whole image and long description code. Then the figcaption tag is used to wrap the long description text. This can be normal HTML text that’s available to both sighted and screen reader users. Inspect the code of this preformatted example to see how the figcaption tag is used house a long description.

Pie chart showing factors affecting credit scores. See table next to this image for the data.

Factors in Your Credit Score

Payment History (35%) 35%
Outstanding Debt (30%) 30%
Credit History (15%) 15%
Pursuit of New Credit (10%) 10%
Types of Credit in Use (10%) 10%

The aria-describedby method:

You can also use the aria-describedby attribute added to an image to point to the id of a long description. For example, you could point to the id of a div wrapping a long description like:

    Pie chart showing factors affecting credit scores. See table next to this image for the data.


Factors in Your Credit Score

Payment History (35%) 35%
Outstanding Debt (30%) 30%
Credit History (15%) 15%
Pursuit of New Credit (10%) 10%
Types of Credit in Use (10%) 10%

Inspect the code of this preformatted example to see how aria-describedby works. In this example, the table within the wrapper div describes the segments of the pie chart. It wouldn’t be possible to descibe all those segments in basic alt text.

The web accessibility guideline SC related complex images is 1.1.1 Non-text Content (Level A). The W3C WAI has a tutorial with more information about handling complex images. Most universities are striving to meet WCAG level AA conformance. Of the three levels of WCAG conformance, level A is the most important to achieve and level AAA is nice to have but not mandatory. So level AA is a middle ground that most universities and government agencies strive to achieve. Ohio State’s information about minimum web accessibility standards can be found in the Digital Accessibility Center’s website.

Comparing screen readers: VoiceOver and NVDA

VoiceOver for Mac and NVDA for Windows are two popular screen readers helping users with visual disabliites navigate websites. They share some similarities and have some differences. VoiceOver’s rotor and NVDA’s Insert+F7 function fullfill a similar purpose.

The basics:

To launch VoiceOver, click Command+F7. To a launch NVDA, click Control+Alt+N.

In VoiceOver you use VO + Right Arrow to read the next item while in NVDA you use the Down Arrow.
In VoiceOver you use VO + Left Arrow to read the previous item but not in NVDA.

The rotor:

The rotor in MacOS VoiceOver is similar to NVDA’s Insert+F7 function in that it allows the user view and to jump to headings, landmarks, or other HTML elements. To popup the rotor in MacOS VoiceOver presss VO + U. ‘VO’ is the VoiceOver keys (Control & Option) that are pressed in conjunction with a VoiceOver command (like U which launches the rotor). Use the Left and Right arrows to select page elements like: Links, Headings, Tables, Frames, Images, Form Controls, Landmarks, Visited Links, and Non-Visited Links. Once a page element type is selected, the Up and Down arrows can be used to select a particular element from the list and Enter can be used to activate it. The Escape key exits the rotor.

An example of the rotor in MacOS Voiceover poping up a list of headings.

An example of the rotor in MacOS Voiceover poping up a list of headings.


Configuring the rotor

You can choose which element types are viewable in the rotor by opening the VoiceOver Utility with VO + F8, then going to ‘Web’ > ‘Web Rotor’. VoiceOver for iOS devices also has a similar rotor that can be configured.



Features similar to the rotor in other screen readers:

Popping up a list of all links in other screen readers is similar to what the rotor does in MacOS VoiceOver. Insert+F7 in NVDA can pop up a list of all links.

To list all links in MS Narrator, hit Caps Lock + F7 or Insert + F7. See Deque System’s information about MS Narrator for info about keyboard shortcuts.

To list all links in JAWS for Windows, hit Insert + F7. See Deque System’s information about JAWS for info about keyboard shortcuts.

ARIA Menu or Menu bar interaction pattern recomendations

Many websites use Menu or Menu bars for their global navigation to pages in the site. For maximium accessbility it’s best for these menu bars to follow ARIA APG guidelines. In ARIA APG’s recommended menu interaction pattern, because menubar and menu elements are composite widgets as described in Keyboard Navigation Inside Components, Tab and Shift + Tab don’t move focus among the items in the menu. Instead arrow keys are used to move through the menu items and the Enter key is used to follow a link. The right arrow goes to the next item and the left arrow moves focus to previous item. See ARIA APG’s recommended menu interaction pattern for more details about how the arrow keys should behave. Using ARIA APG guidelines as a standard way of building menus will help establish a web convenion that users can learn and apply to a lot of website. Using a standard interaction pattern convension for menus will save users time in not having to learn a different interaction pattern.

ARIA menu and interation pattern resources

What’s new in WCAG 2.1?

The Web Content Accessibility Guidelines (WCAG) are a set of globally adopted technical standards on web accessibility developed by the World Wide Web Consortium (W3C). The guidelines represent a shared, international standard developed by many different stakeholders including: education, industry, disability organizations, government, and accessibility research organizations. The guidelines strive to make websites, apps, electronic documents, and other digital assets accessible to people with a broad range of disabilities, including sensory, intellectual, learning, and physical disabilities.

WCAG covers a breath of success criteria that, when followed by website designers, developers, and content authors, can remove many of the barriers that people with disabilities face when navigating a website or other digital applications. Find out more about WCAG in this article – WCAG 101: Understanding the Web Content Accessibility Guidelines.

What are the new accessibility guidelines in WCAG 2.1?

WCAG 2.1 became a W3C Recommendation in June 2018. There are 17 new SC (success criteria) guidelines in WCAG 2.1:

  1. 1.4 Reflow (AA) – responsive design principals are important to accessibility.
  2. 1.4.11 Non-text Contrast (AA) – requires the icons, UI, and graphics should have good color contrast.
  3. 1.4.12 Text Spacing (AA) – the Text Spacing bookmarklet can help test.
  4. 1.4.13 Content on Hover or Focus (AA) – don’t show unexpected content on hover.
  5. 2.1.4 Character Key Shortcuts (A) – don’t interfere with screen reader shortcuts.
  6. 2.5.1 Pointer Gestures (A) – complex gestures also need a single pointer equivalent.
  7. 2.5.2 Pointer Cancellation (A) – down events must not be used to complete a function.
  8. 2.5.3 Label in Name (A) – visible labels should match the accessible name.
  9. 2.5.4 Motion Actuation (A) – don’t make shaking the phone the only way to complete an action.
  10. 4.1.3 Status Message (AA) – alert the user, but don’t interrupt them.
  11. 1.3.6 Identify Purpose (AAA) – HTML content should provide context, purpose and meaning.
  12. 2.2.6 Timeouts (AAA) – you must store the user’s data for 20 hours.
  13. 2.3.3 Animation from interactions – motion animation trigged by interaction can be disabled. Use media query: prefers reduced motion.
  14. 2.5.5 Target Size (AAA) – touch target size must be 44 x 44px.
  15. 2.5.6 Concurrent Input Mechanisms (AAA) – websites should allow users to switch between input modalities.

Here’s a Good video covering all things new in WCAG 2.1 (53:58).

What are the main WCAG principals of accessibility?

The top four principals of WCAG are: perceivable, operable, understandable, and robust. All new guidelines introduced fall under one of these broad categories. WCAG 2.2 is scheduled to be completed in early 2023 and will include 12+ new guidelines.

  1. Perceivable – people must be able to find your content.
  2. Operable – people must be able to use your website.
  3. Understandable – people must be able to understand your website.
  4. Robust – your website must work with different assistive technology.

See Understanding the Four Principles of Accessibility for more info.

Setting up a heading structure in InDesign for accessibility

Something that’s different in InDesign from Word is that in MS Word you already have predefined Heading level styles in the Home ribbon whereas in InDesign you have to set up your own ‘Paragraph Styles’ and map them to a Heading structure. You can follow the directions found in this workflow for creating accessible PDFs from InDesign article to setup a template InDesign document with a structure. Here’s an example template exported to PDF. For an InDesign template for graphic designers, it might be good to have Paragraph Styles predefined and maybe have an XML document structure predefined based on text frames they most commonly use. Penn State also has some good information about accessibility and InDesign in this article: Accessibility at Penn State | InDesign Accessibility (

What you setup in the InDesign Structure panel becomes what you see in the Tags panel in Acrobat. The tags in PDF are very similar to HTML with a few odd differences. For example there’s a general <L> tag for all lists instead of <ul> and <ol> (unordered lists and ordered lists) tags. InDesign can also export to ePub format which is more like regular HTML.

Set the tab order in InDesign:

The structure created in the InDesign Articles panel is exported to the PDF file, but it is not automatically set as the order followed when tabbing through interactive elements like hyperlinks and form fields.

  • If the tab order fails in Acrobat, you can right-click on the issue and choose ‘Fix’. This sets the tab order to match the document structure that you set up in InDesign.
  • You can view and adjust the reading order in the Order panel (looks like a Z icon in the left panels).
  • If the panels aren’t there you can open them by going to: View > Show/Hide > Navigation Panes > Page Thumbnails
view and adjust the tab order in the tab order panel

To get to the Tab Order Panel in Acrobat, click the Z icon in the left panels.


How to Export to PDF:

In InDesign, when you are ready to export to an accessible PDF, go to: File > Export > PDF

  • Choose Adobe PDF (Interactive) if your document contains forms, buttons, audio, or video or Adobe PDF (Print) format if it doesn’t.
  • Check the Tagged PDF box to ensure you document structure and tags are exported.
  • Then save the document and open it in Acrobat Pro to check for any accessibility issues.
Export to PDF settings - create tagged PDF

In the export PDF settings in InDesign, check the Tagged PDF box to ensure you document structure and tags are exported in the PDF output.


Three things to do in Acrobat after you export your InDesign document to PDF:

  • Assign a document language
  • Set PDF tab order to ‘use document structure’
  • Run the accessibility checker
  • View and adjust the order in the Tab Order Panel
view and adjust the tab order in the tab order panel

Screen of the Tab Order Panel in Acrobat where you can adjust the PDF document tab order.


Find out more in the workflow for creating accessible PDFs from InDesign article.

HTML 5 Landmarks and equivalent ARIA Landmark Roles

WAI-ARIA roles provide more semantic meaning to content making webpages more accessbile. ARIA landmark roles provide a way to turn bare <divs> into HTML 5 landmarks. This can be useful in systems where the <divs> can’t be turned into real HTML 5 landmarks for whatever reason.

Comparison of HTML 5 Landmarks and equivalent ARIA Landmark Roles
HTML 5 Landmark Equivalent ARIA Landmark Roles more information
<main> <div role="main"> info about the ARIA main role
<nav> <div role="navigation"> info about the ARIA navigation role
<header> <div role="banner"> info about the ARIA banner role
<footer> <div role="contentinfo"> info about the ARIA content info role
<section> <div role="region" aria-label="Example"> info about the ARIA region role
<aside> <div role="complementary"> info about the ARIA complementary role
<form> <div role="form" aria-label="Contact information"> info about the ARIA form role

Find about more about ARIA (Accessible Rich Internet Applications markup) in the Custom JavaScript/ARIA Widgets Buckeye Learn course.

Fixing a broken skip nav accessibility error

The Ohio State navbar contains a hidden skip nav link that jumps to an anchor called “page-content”.  This skip link is hidden visually (through CSS) but available for those using screen readers to use to skip the other Ohio State navbar links and go straight to a page’s content.  A review by the WAVE tool will report if the skip nav anchor is missing from a page. To correct a broken skip nav link, add an anchor near the start of your page’s main content. Here are some examples of possible anchors:

  • <a name=”page-content” id=”page-content”> – The ‘name’ attribute works in HTML 4 and the ‘id’ attribute in HTML 5.
  • <div id=”page-content”>  – In HTML 5 the ‘id’ attribute can be placed in a lot of different elements, such as a <div> or <main>.

Some more information about using anchors in HTML is available from W3 Schools. See the info about the HTML <a> name Attribute.

The WebAIM organization has some more information about “Skip Navigation” Links in their site.