Picture lots and lots of dogs wearing vests
Now, are you thinking of a Lewis Carroll-esque canine tea party, or a collection of service dogs? If the latter, you may be ready to attend the thankfully-abbreviated California State University: Northridge Assistive Technology Conference (hereafter referred to as CSUN or #CSUNATC18).
If you picture an impeccably-dressed doggy social occasion, you would also be ready to attend CSUN. In fact, pretty much everyone is welcomed at the conference: it is easily one of the most diverse and inclusive tech communities out there. In starting conversations with anyone who would sit still long enough, I met people who had been working on web accessibility for decades and contributed to the W3C, and also people entirely new to the field getting their introduction at the conference itself. The one connecting thread (besides an interest in accessibility) was how uniformly friendly and approachable both attendees and speakers were.
Based on a quick scrape of the 2018 session list, there were 465 session offerings spread over three days. It was only possible to sample a small fraction of the available content, but I have grouped a selection of the sessions I did attend into the following categories:
- What even is accessibility?
- How to talk to people about accessibility
- The nitty-gritty: how to implement proper accessibility
- New & shiny tech
For each presentation below, I have provided a summary of my takeaways along with links to slide decks or other presentation-related material where available. There is a more comprehensive list of available presentations on Github put together by Jamal Mazrui.
The defintion of accessibility is surprisingly difficult to pin down, even after working in the field for multiple years. It’s pretty easy to come up with an abstract definition, something like “access for everyone regardless of ability, age, gender, country, etc.,” but difficult to understand what that means for any given project. CSUN talks came to the rescue with both highly specific legal ways to define accessibility, as well as UX-oriented rules of thumb.
2018 Legal Update
Lainey Feingold • slide deck
This session was probably the most popular presentation of the conference, offered in three separate time slots. Lainey Feingold addressed the legal angle of how web accessibility is interpreted in the U.S. and went over recent and pending cases. The short version is: conform to WCAG 2.0 AA; even if you’re not the government, the ADA probably applies.
If anyone doubts the efficacy of legal action, I saw representatives at CSUN from multiple companies whose lawsuits get mentioned in this presentation.
The State of Accessibility, 2018
While Lainey focused in depth on the legal meaning of digital accessibility in the United States, Preety Kumar gave a fascinating overview of how some other countries legislate and enforce access. Many other nations also directly use WCAG as a legal basis, but I was particularly interested in two that don’t: China and India. China currently has a web accessibility standard, but compliance is “voluntary”, and India has a mandatory “Rights of Persons with Disabilities Act” but no regulatory body for compliance testing.
Preety also made an interesting point that average website accessibility as measured by automated testing varies quite a bit by industry: retail sites have received quite a few lawsuits recently, and test failures across the sector have decreased; on the other hand, financial sites have accumulated more failures in the past year. Hopefully, Preety can make her slides available online, as they are densely packed with interesting data.
Caitlin Geier and Denis Boudreau • powerpoint slides
Along the same vein as Jakob Nielson’s 10 Usability Heuristics for UI Design article from 1995, Caitlin Geier and Denis Boudreau put together ten accessibility heuristics as reasonable rules of thumb in any project. These are most useful in the planning and design stages, but would still be a good reference for anyone. There’s a PDF with detailed explanations and examples, but here’s the short version:
- Navigation and wayfinding
- Structure and context
- Color and legibility
- Language and readability
- Error prevention and error states
- Predictability and consistency
- Alternatives to audio/visual content
- Account for multiple interaction methods
- Provide enough time and preserve information
- Control over movement and flashing
Trickle Down Accessibility
Ted Drake and Sarah Margolis-Greenbaum • summary and slides
Ted Drake and Sarah Margolis-Greenbaum challenge the assumption that web accessibility can be accomplished by focusing on screen readers. They went over some of Intuit’s user testing and lessons learned, effectively driving home the point that focusing only on vision impairments is not enough.
A couple of specific points from the presentation stood out to me:
- Improving readability through layout and simple language is one of the easiest ways to improve access to a large number of people.
- Spending time designing for disabilities helps everyone. For example, Intuit added the ability to search for accountants by languages spoken because a deaf pizzeria owner wished to find someone who could communicate with ASL. Now anyone can filter search results by any language.
A11y Wars: Rethinking Interpretation Differences
Glenda Sims and Wilco Fiers
Unclear expectations and poor communication can send a good-faith effort into a spiral of arguments about what constitutes “good enough” accessibility. Glenda Sims and Wilco Fiers suggest splitting recommendations into three categories, and recognizing that there are valid use cases for each one:
- Minimal: urgent fixes with the minimum effort to fix broken UX
- Optimized: effort is put in to follow best practices and fulfill the spirit of WCAG, rather than simply meeting success criteria.
- Ideal: pushes the boundaries of what is possible, and strives to provide a smooth experience for all assistive technology.
However tempting it is to always push for optimized or even ideal accessibility, sometimes minimal compliance makes more sense – for example, pushing quick fixes to a legacy project nearing its end of life. Recognizing and accepting that different situations call for different approaches will hopefully make communication and feedback less stressful. At the very least, everyone involved can be explicit about which level they’re targeting and sort out differences before starting in on implementation.
Everything I Know About Accessibility I Learned from Stack Overflow
Adrian Roselli • slide deck
The title of this talk is a little misleading – Adrian Roselli very clearly did not learn accessibility from StackOverflow, but he does spend time on the site answering questions about accessibility. The selection of questions he highlighted was absolutely hilarious but emphasized the broader point that many (if not most) developers are isolated from the accessibility community and sharing information that may be misleading or outright wrong.
As things stand right now, StackOverflow remains in a bubble, as evidenced by its recent developer survey highlighting the notable absence of people with disabilities (among other groups cough women and people of color cough). Although Adrian focused on StackOverflow, it isn’t the only echo chamber.
Most of the sessions fell into this category and ranged from high-level blueprints for accessibility programs to specific technical recommendations. These are my favorites from among those talks:
Today’s Most Challenging Accessibility Obstacles
Hans Hillen • powerpoint slides
The format of this talk leaned more towards some mutually enjoyable back-and-forth grumbling about common antipatterns between speaker and audience. All of the examples were relatable, but some were particularly enjoyable to vent about:
- placeholders replacing form labels
- hijacking scroll
- text over background images (and the inherent contrast problems)
Hans Hillen also identified cases where there still is no good and easy solution: responsive data tables, context changes in single page apps, and complex non-native components like autocompletes and date pickers stood out the most.
Managing Focus within Data Tables
Elizabeth Whitmer and Pavani Gonuguntla • summary and demos
Sometimes a deep dive into a specific problem can be as useful as a high-level overview. This talk detailed the decisions made while creating a large and fully-featured data table. Particular areas of difficulty included handling focus when sorting, adding and removing rows and notifying non-visual users of updates. It is well worth it to take a close look at their demos.
Automated and manual accessibility testing
Apparently, I cannot get enough of accessibility testing, because I went to three separate talks on different aspects of it. The big takeaways centered around using the right testing approach for the right task.
Automated testing will provide scalable, consistent feedback but will never be able to catch all accessibility issues. Manual testing can produce much more comprehensive results and captures the human element, but is not scalable, can be inconsistent, and is not reusable.
Even within automated testing, there is a divide in which issues unit tests may capture and which need integration testing. Some issues that depend on integration tests include inter-component focus management, color contrast, content structure, and language. The talks also mention a good resource for learning more: the W3C auto-wcag community group dedicated to creating guidelines for automated accessibility tests.
The talks in this category leaned much more technical, and are likely to be more helpful for someone who already has a good handle on the basics and planning for the future.
The trade show is also relevant to this section – among the booths upon booths of new or updated assistive tech, captioning services, consulting agencies, and giant tech companies, there was the occasional genuinely mind-blowing project. One such piece of tech was the Japanese Company QD Laser’s retinal imaging glasses. They use a front-mounted camera and low-powered laser to project perfectly sharp images directly on your retina. I was able to try them and can confirm I saw a square video like a semi-translucent overlay on reality. It even comes with a zoom-and-enhance ability.
Accessibility Object Model (AOM)
Alice Boxhall and Dominic Mazzoni • slide deck
AOM is a proposed programmatic API for modifying the accessibility tree in the same way that the DOM allows modification of the document tree. For example, instead of needing to declaratively set something like
<input aria-labelledby="foo">, you could simply do
inputElement.ariaLabelledby = fooElement;(not necessarily the eventual syntax). AOM would make possible a whole host of nice enhancements, but two of my favorite are these:
- Default semantics for custom web components
- No more
ids sprinkled everywhere
Controlling Your Computer with Your Eyes
Apart from being absurdly cool this isn’t directly relevant to web design but is instructive in showing a little-considered method of interacting with websites. In the demo, Harish was able to move the cursor, zoom to refine the cursor position, click, and type using a virtual keyboard. One more reason to shy away from tiny buttons.
ARIA 1.1 in-depth
Suman Damera and Denis Boudreau • slide deck
We’ve already been using a few of the ARIA 1.1 attributes covered in this talk –
aria-haspopupto name a couple – and the others look just as useful! Suman and Denis’ slides have an excellent summary of 10 of the new properties, well worth checking out.
Working with accessibility can sometimes feel isolating since the community is relatively small, but it doesn’t feel that way at CSUN. The best part of attending the conference is being surrounded by other people who genuinely care about accessibility and spend their time thinking about it. It’s the perfect place to get a broad sense of what people in the field are working on, and to gather inspiration for your efforts.