Apple’s Inaccessibility

Apple has traditionally been a great advocate and model for accessibility and technology. Unfortunately this hasn’t been the case lately. One could even argue that default settings and recent designs are even counterproductive to accessibility progress. This includes VoiceOver, keyboard access, and design decisions.

Bug Infested

To begin, let’s reference a recent article by Marco Zehe where he explains major VoiceOver bugs in OS X 10.10 (Yosemite) and iOS 8. The two major examples he cites are:

  • When VoiceOver is running on the iPhone, using the Back button (or Scrub gesture to return to the previous page) will freeze VoiceOver.
  • When VoiceOver is running on the iPad, using Safari and the use of WebView components trigger app crashes and OS restarts.

In addition to bugs, I’ve noticed the following blatant accessibility problems pop up in Apple’s products. Each of these alone may not be a showstopper but collectively it shows a pattern of a company that doesn’t care, or pay much attention any longer; a company that’s losing its edge.

VoiceOver Hints

By default, VoiceOver places a large delay when announcing “hints”; this creates a huge lag for the reading of text defined by the aria-describedby attribute. And we all know that the support of aria-describedby is becoming more and more prevalent and essential in today’s world of modern web development. The delay is so long that it creates confusion; developers and testers very often think something is broken.

To “fix” the setting, you must find the deeply buried option and modify the delay time; in System Preferences, go to Accessibility / VoiceOver / VoiceOver Utility / Verbosity / Announcements (in OSX 10.9, it’s the the last setting).

Keyboard Access

A great example of Apple’s inaccessibility is the setting for keyboard focus; by default it doesn’t allow for typical keyboard navigation! A user cannot use the tab key to access all controls in an interface. In order to fix, one must again, search for the appropriate settings and modify. Again, this can be confusing and frustrating for developers and testers, let alone regular users.

To fix, the first step is to go to System Preferences / Keyboard / Shortcuts, and in the last section “Full Keyboard Access”, ensure the radio options “All Controls” is selected.

Let’s refer to the follow articles for more details on how to resolve which includes specific browser settings:

Design Problems

There are two major design issues by Apple recently which hinders accessibility: animations and parallax effects, and flat design.

There are many vestibular-related issues in Apple’s design, most notably during the release of iOS7. The issue continues to iOS8 although iOS settings are now available to help resolve the issue. To reduce parallax-type effects in iOS, go to Settings / General / Accessibility / Reduce Motion.

Also, iOS flat design (and the trend in general) is bad for usability and accessibility. Mostly because flat design creates ambiguity between elements; as the Nielsen Norman Group report states, “flat design hides calls to action“. And by implementing flat design, Apple indirectly encouraged others to do the same.

In addition to Nielsen Norman Group’s finding, another usability expert Steve Krug agrees that flat design is a detrimental practice. Here’s a quote from his book which was tweeted last May:

Hope and Warning

Let’s hope Apple returns to the practice of releasing quality products: everything just works and accessibility is continually improving. And fair warning, after the mobile wars a few years ago, Google tops Apple in mobile operating systems; will accessibility be next?

Related Reading

2014 Year in Review

2014 was surely a much busier year than expected. It started a bit slow, but sure got busy!

In the most recent blog post, the hot topic of Google’s new version of reCaptcha dubbed “No Captcha” was addressed. Although there are remaining challenges, Google’s No Captcha Shows Some Progress.

In a guest post by Jennison Asuncion, a new date for Global Accessibility Awareness Day (GAAD) was announced. It’s now the third Thursday of May.

In the post Floated Labels Still Suck, problems and fixes are discussed for the terrible design pattern of putting input labels inside input fields.

Great progress for accessibility continues in WordPress; a podcast with two WordPress contributors, Joe Dolson (@JoeDolson) and Joseph O’Connor (@accessibleJoe), was published in September.

Web Axe author Dennis Lembree read Steve Krug’s excellent book, “Don’t Make Me Think, Revisited” and Twittered a series of accessibility-related points. The series was published in the post Tweets quoting “Don’t Make Me Think, Revisited”.

In May, I announced that Easy Chirp now provides accessible images for Tweets. This feature is badly needed and isn’t available on any other Twitter app. Unfortunately, and surprisingly, the feature is grossly underused.

In March, we posted a summary of CSUN14, the 29th Annual International Technology and Persons with Disabilities Conference. It was another great event; thanks again to California State University, Northridge. (Look for Dennis at CSUN15!)

On a more personal note, Web Axe author Dennis Lembree released an open-source Accessible HTML5 Video Player in September via his day job at PayPal. He recently changed jobs and is now Product Manager, Accessibility at eBay.

More from 2014:

Google’s No Captcha Shows Some Progress

It sure was exciting when Google’s new reCaptcha was announced last week. Dubbed “No Captcha”, the goal of course is to provide a service that determines a human from a bot in order to prevent spam and abuse of online forms.

Derek Feathersone (@feather) from Simply Accessible was one of the first to report its accessibility impact in the blog post On the Accessibility of Google’s No CAPTCHA. The tone of the post is very positive, but the testing cited excludes JAWS with IE (still the most popular screen reader combination) where I found the No Captcha failed miserably. On the bright side, it passes with keyboard-only, Dragon Naturally Speaking, NVDA and the latest VoiceOver.

Two fundamental problems with No Captcha is that it requires JavaScript and it doesn’t work with touch devices.

You can try the No Captcha yourself on a test page hosted by Alastair Campbell (@alastc).

checkbox with label "I am not a robot"
Google’s “No Captcha” reCaptcha

At best, No Captcha simplifies the Captcha experience. At worst, it excludes some users even more than the previous version. Hopefully Google will fix the current issues, especially support for JAWS.

My recommendation is to continue using non-Captcha security techniques; two great articles on how to do so are Spam-free accessible forms by WebAIM and 10 Things to Check Before Using a CAPTCHA by SitePoint.

Related:

Announcing GAAD’s New Date for 2015 and Beyond

This is a guest post by Jennison Asuncion.

GAAD logo

When Joe Devon and I launched Global Accessibility Awareness Day in 2012, neither of us had any idea if it would take flight. It still honestly amazes us how much the mainstream tech and the digital accessibility communities, in addition to others who have an interest in raising awareness of digital accessibility and inclusion have run with the idea. As of 2014, in-person and virtual events, those targeted at the general public and internal audiences have been held on six continents and in multiple languages. We truly appreciate all the support and eagerly await what everyone has in store to celebrate GAAD in 2015 and beyond.

May 9 was the date that Joe and I randomly chose when we first launched GAAD, and we marked the event on that date in 2012 and 2013. When we were advised this date conflicted with some local holidays, this year, we moved it to May 15. After further discussion, and to try simplifying things, Joe and I have decided that rather than a specific date, starting in 2015, we invite you to join us in marking Global Accessibility Awareness Day on the third Thursday of May. This means that the next GAAD will fall on Thursday May 21, 2015.

“we invite you to join us in marking Global Accessibility Awareness Day on the third Thursday of May”

Thinking of marking GAAD? Take a look at some of the public events and other activities from the last three years. To stay up-to-date, Like GAAD on Facebook or follow GAAD on Twitter. e-Mail us with any questions at globala11yawarenessday at gmail dot com.

Floated Labels Still Suck

Back in June of 2012, I wrote that the Placeholder Attribute Is Not A Label. The post points out that it’s bad practice to use text to look and behave as a placeholder rather than a label. This is generally known as “floating labels” or “inline labels” (I also call this LAP for short, “labels as placeholders”.) And as you may know, labels are required for accessibility (and usability) as they identify to the user what the input is for. It’s a general best practice to provide a label for every form input.

Unfortunately this design trend still continues even despite repeated concerns on use of placeholder from web development and accessibility experts. Web Axe too maintains that this pattern is bad for a variety of reasons which weren’t fully listed in the original post.

Example of labels acting as placeholders
Don’t do this!

Sources of the Problem

Like many accessibility problems on the web, floated labels can be attributed to multiple issues. Firstly, web designs are trendy, and when a huge well-known company does something like floated labels (i.e. Apple), others follow, even if it’s a poor practice. And if usability testing is done and includes floated labels, it may not provide reliable results because the usability testing most likely:

  1. is not done with long forms (which greatly increases the needs reviewing input and increases the potential of losing context)
  2. is not done with users with disabilities
  3. is not done users with a situational disability
  4. is done most likely in a controlled environment (latest browser provided, no chance for JS delays nor errors, guided by moderator, not a real product)

The web development community has a lack of adequate HTML skills which also contributes to this problem, so bad that articles recommending basic HTML practices must be written, and frankly, are much needed. A fundamental HTML practice is that a text input requires a label to define what it is. The intent of the placeholder attribute is very often misused; the W3C gives an explicit warning about placeholder:

Use of the placeholder attribute as a replacement for a label can reduce the accessibility and usability of the control for a range of users including older users and users with cognitive, mobility, fine motor skill or vision impairments

List of Pitfalls

Whatever the reason floated labels exist, here is a list of potential issues with the pattern.

  1. The simple fact that since the label disappears when entering text, the user may forget what the input is for. This is especially important when reviewing form data; for users with a cognitive accessibility; and people who may be distracted when completing a form.
  2. This pattern can’t support both a label and a placeholder.
  3. Depending on user’s environment and implementation, there may be a delay in rendering/performance: on page load and in text disappearing after beginning typing. Or even it breaking entirely; see the two examples below.
  4. If an HTML5 placeholder is used for label, older browsers and/or assistive technologies may not support it at all. For example, there is no placeholder support in IE9 and Opera Mini.
  5. Placeholder text is rendered differently across browsers/versions; extra design and development is required to make them accessible (color contrast is usually very low) and have consistent styling. The traditional label is straight-forward and reliable.
  6. Developing this pattern in the best way possible can be fairly complicated and is inconsistent among implementations. Code becomes heavier, more fragile, and more difficult to maintain.
  7. The pattern doesn’t support standards; it doesn’t implement HTML elements and attributes as intended. The W3C HTML5 placeholder specification specifically states it should be a “short hint” of user input and states: The placeholder attribute should not be used as an alternative to a label.
Placeholder! Why U no use label?
A great meme!

In the Wild

No matter how many issues exist, the fact is that the floated label pattern is prominent on the web. Here are a couple of unique examples of this pattern in the wild which have completely gone wrong. This demonstrates only one of many issues with this pattern; sometimes the label doesn’t disappear visually and the label is blocking the view of the input text.

Hootsuite sign-in form; the password input contains both the floated label and the masked password, making it unreadable.
Hootsuite sign-in fail.
In Facebook app, text input for birthday message is blocked by floated label.
Facebook birthday fail in iOS app.

Attempts to Fix

Attempts have been made to help address or fix this problem, but not with much traction. One idea is to ask browsers to darken the placeholder text so that it passes WCAG color contrast requirement. This is a plausible suggestion but not likely to happen, especially any time soon. See below for the CSS fix. And doing so only addresses one issues; it wouldn’t fix the fundamental problem of the label disappearing.

Another attempt is to, upon focus of the input, reduce the text size of the floated label and move it to the bottom of the input field (rather than it disappearing entirely). This is not a feasible solution either as this behavior may cause confusion, requires text size that’s too small, and still adds code weight. Besides, if the label is going to remain visible near the input, it should just be there in the first place.

How to do it right

To create a form that’s robust and accessible, use the label and placeholder as intended:

  • The label element represents a caption in a user interface. It tells the user what the purpose of the input.
  • The placeholder attribute represents a short hint (a word or short phrase) intended to aid the user with data entry when the control has no value.

In the example below, the label “Enter URL” clearly indicates what the text input is for. The placeholder “http://” is a short hint as to what text should be entered. Also, a good guideline for placeholder text is similar to title attributes: only use them for supplementary information, not required information.

A text input with label "Enter URL" above the input and http:// as placeholder value.
Example of recommended use of input label and placeholder.

An exception to always displaying a visible label may be the search input where an icon (such as a magnifying glass) and the location of the form on the page visually denotes it’s meaning sufficiently. A visually hidden label is still required.

To better support mobile devices and screen enlarging tools, place the label above the input (as opposed to the left).

If using a placeholder (hopefully for the right reason), use CSS to darken the light grey so that it adequate color contrast. Here’s an example:

::-webkit-input-placeholder { /* WebKit browsers */
    color:    #666;
}
:-moz-placeholder { /* Mozilla Firefox 4 to 18 */
   color:    #666;
   opacity:  1;
}
::-moz-placeholder { /* Mozilla Firefox 19+ */
   color:    #666;
   opacity:  1;
}
:-ms-input-placeholder { /* Internet Explorer 10+ */
   color:    #666;
}

Summary

Please don’t use “floated labels” or “labels as placeholders” technique; it can break the user experience for a large variety of reasons. Use the label element and placeholder attribute as intended. Use CSS to ensure placeholder text has adequate color contrast.

Further Reading