Standard Practices

It was a cloudy day in June, cool for the season, and I was trying to read my email, though “read” had taken on a meaning I had never imagined. I held my phone to my ear as it spat out quick successions of dates, senders, and subjects—I was searching for an email from my doctor, who was supposed to send me some paperwork regarding a serious concussion I had sustained three days earlier. My screen reader was finally working, and the flood of information washed over me: fast, semi-comprehensible, completely overwhelming. Around the tenth email, I noticed a pulsing in my field of vision. I put the phone down, burying my head in my knees as the migraine set in.

Like most people who wind up working on web accessibility, I had a catalyst: while recovering from that severe concussion, I spent a year using screen readers and speech-to-text software to do even the most basic digital tasks. After spending my whole life navigating the digital world with sight, touch, hearing, cognitive acuity, and full mobility, I was having trouble adjusting to the brittleness of the Internet when I could no longer see it, to how fragile our digital world proved to be when just one sense was stripped away.

That was eight years ago. At the time, I’d assumed that the issue was that accessibility online wasn’t standardized, or maybe it was just too complex. After all, my screen-reading software seemed to glitch on every page, reading out excess information that I didn’t need or couldn’t parse. What I didn’t know then was that a standard did exist—and had existed for over 25 years—while I struggled to read an email. That standard is WCAG: the Web Content Accessibility Guidelines.

The Promise of Standards

The World Wide Web Consortium (W3C) released the first version of the Web Content Accessibility Guidelines in 1999. “It has always been difficult to know, when making a site more accessible, which changes are critical,” said then-W3C director Tim Berners-Lee in the accompanying press release. “These guidelines answer that question, and set common expectations,” he explained, noting that some of these guidelines would be irrelevant once tools for authoring HTML could do them automatically. The first major standard for web accessibility had been set, and it was relatively easy to meet—technologically speaking.

Berners-Lee’s pronouncements seemed to herald greater progress in web accessibility than ever before. His reflections on the guidelines’ first iteration are tinged with the heady optimism of the early web: with the right set of rules, progress would come naturally, advancing almost without effort. Rules would beget tools, and tools would render rules obsolete—or so the theory went. That was how it had gone, after all—the right rules about HTML or the URL specification had facilitated the rise of the early web. Their adoption was critical to the Internet’s development.

WCAG was the latest standard from the four-year-old W3C—a consortium of academics and major players in the industry: companies like Microsoft and IBM and Netscape and Sun Microsystems, who sent money and representatives to help develop the standards for everything that made the early World Wide Web into today's Internet. Unlike purely technical specifications, WCAG addressed multiple access modalities—what were then called "user agents," encompassing everything from keyboard navigation to screen readers. The guidelines took the form of checklists with three priority levels (A, AA, AAA), allowing developers to display nifty compliance badges from the W3C website as proof of their accessibility efforts.

Thanks to the cultural capital of the W3C, WCAG has enjoyed near-monopoly status as an arbiter of digital accessibility—no other comprehensive set of guidelines exists. Every regulatory framework that has adopted digital accessibility finds itself, willingly or not, entangled with WCAG's specifications. WCAG’s promise, steeped in 1990s disability rights discourse and the optimism of the early web, is that anyone trying to access online content, by any modality, should be able to do so with minimal friction. Whole industries of conformance and compliance have developed around it; every basic accessibility guide departs from the building blocks it sets.

WCAG and I grew up together—both born in the late '90s, both shaped by the promises of the young, open Internet. Yet eighteen years after WCAG first came out, I couldn’t do something so rudimentary as efficiently finding a specific email without looking at my screen. In 2025, the accessibility advocacy organization WebAIM reported that 94.8% of home pages contained an average of 51 WCAG errors per page. Despite its widespread adoption, WCAG isn’t having the effect that Berners-Lee predicted.

After a stint in industry working on accessibility, I wanted a better answer as to why. I spent four months interviewing web accessibility professionals. These people are testers, developers, product managers, program managers, engineers, or all of the above. They work with WCAG day in and day out, using their professional positions to advance accessibility. All their work occurs because of WCAG and its corresponding regulation—but their relationship with the guidelines ranges widely.

Meaningful markup: HTML as a basis for web accessibility

When I first met Jordan,* they introduced themself to me as a web developer. They’d spent a few decades as a web developer across the stack, not exactly a newcomer to the field. Despite these many years in the industry, Jordan had only ever heard about web accessibility after their company got sued for being inaccessible. Jordan’s team, which was working on a design system for the site, was put on the job.

“The way I learned HTML, in high school in the early 2000s, [...] is that these are the tags that make the text look a certain way,” they told me. “But over the years a lot of really smart people have worked on putting more meaning into the HTML itself and giving us the tools to express content in a richer way.” As a result, Jordan said, HTML provided tools for adaptable user experiences, and even laid the groundwork for an Internet that could be readable by machines. At a time when complex frameworks dominate development, it was refreshing to hear someone extol the virtues of pure, simple HTML. Many web accessibility practitioners spend a great deal of time doing just that: discussing semantic HTML, the foundation of contemporary web accessibility.

That’s no accident. The foundations of accessible web design stretch back to 1971, when Gregg Vanderheiden founded the TRACE center at the University of Wisconsin-Madison to research technology for people with disabilities. By the 1980s, TRACE was collaborating with major tech companies, influencing industry standards while shaping the technology itself. By 1987, their accessibility work had reached millions through Mac OS. The momentum built slowly, deliberately, systematically. By the time the web emerged, TRACE was the first to publish a set of accessibility guidelines in 1995. This first version had only six areas of focus, each of which had both “Solutions today” and “Solution strategies in the future”—web accessibility presented such fundamental challenges that even basic issues remained unresolved.

Though brief, these early guidelines were vital in shaping HTML's development. Through persistent advocacy, disability considerations became architectural decisions. Many semantic elements we use today like <nav> and <main> exist specifically to help screen readers navigate content without additional markup. These aren't specialized accessibility features but core HTML elements everyone uses. Because its development was influenced by disability considerations from the very beginning, today's HTML is structured to support accessibility.

Machines, Accessibility, and Universal Design

A few months before the TRACE guidelines’ release, Tim Berners-Lee made note of web accessibility in his plenary at the 1994 WWW conference. Most of his talk was not actually about web accessibility at all—instead, it focused on improving HTML markup. The dream of a world wide web involved the ability to find information programmatically, which meant building web pages so that they could be read by machines. One of Berners-Lee's slides displays one view of a webpage with limited markup, as “seen” by a machine. Then, with enhanced markup, the same page reveals its hidden structure. Suddenly, the people, places, and ideas came to life on the page. The web transformed from simple text into a living network of relationships that machines could understand, simply by making markup more explicit and descriptive.

I was startled at first to see web accessibility and machine readability in the same presentation, but it makes sense that the origins of web accessibility were being discussed in the same breath as automated information retrieval. Screen readers, after all, are machines themselves. This convergence echoed prominent discourses of disability in the 1990s: that “good design” could be used by people with disabilities as well as by “regular”—that is to say, nondisabled—people, and that by designing with disability in mind, technologists could create better products “for everyone.” This philosophy is called Universal Design, and it gained popularity following the passage of the Americans with Disabilities Act (ADA) in 1990.

The ADA was a landmark piece of legislation, and few other countries have a law like it. It took decades of disability activism to attain, beginning with protests in the 1970s demanding that the government enforce existing legislation that required entities receiving federal funding to provide equal access to people with disabilities. In March 1990, activists with disabilities cast aside their mobility aids and crawled up the steps of the US Capitol to protest its inaccessibility in an act of protest called the Capitol Crawl. Four months later, the ADA became law, enshrining the right for people with disabilities to have “reasonable accommodations” in any facility, public or private, open to the public.

Universal Design arose as a way to design, from the beginning, with "everyone" in mind, thus avoiding the need for accommodations in the first place. It propagated the idea that designing for disability was simple, universally beneficial, and just good business. The prototypical example is the ramp or the curb cut—unlike steps, they enable anyone, whether on wheels or on foot, to access an elevated platform or sidewalk, creating a common pathway that accommodates different abilities. “Curb cuts” have become a shorthand for disability-focused interventions that wind up having wide-ranging benefits. One such example from web accessibility is the alt attribute on images, which allows users to embed an image’s textual description, also known as “alt text,” directly into the tags of the HTML <img> element. Because alt text makes it easier for web crawlers to identify site information, adding alt attributes to images became a common strategy to boost search engine optimization, making more sites more accessible to screen reader users. This increased popularity had yet another surprising consequence: it created a boom in easily-scrapeable pairs of images with their text alternatives, which became a vital resource for computer vision research that teaches machines to “see.”

Universal Design was and still is a productive philosophy, encouraging companies to design with disability in mind, by pushing them to view designing for disability as a cost-free optimization. This simplicity was its own deception. Truly accessible design confronts complexity at every turn. It must accommodate varied disabilities. It must reconcile contradictory needs. It requires specific attention to the needs of disabled populations, not just a blanket “design for everyone” attitude. By positioning design for disability as “just good design,” Universal Design defanged disability politics while reframing disabled people as just another market segment.

For the creators of the early web, however, Universal Design meshed with the idea of a technology that could be open and flexible to anyone, anytime, anywhere—the very dream that defined the World Wide Web. Today, the W3C’s page on web accessibility describes the web itself as a technology that is universal to everyone, “whatever their hardware, software, language, location, or ability.” That description of the Internet would not have been so prominent without the influence of people like Vanderheiden in the 1990s, whose vision of Universal Design fundamentally altered the idea of what the Internet should be, and paved the way—or, more precisely, the ramp—for new ways of using machines to read and see the digital world.

The Double Life of WCAG

The late 1990s were a strange time on the early Internet: though framed by the libertarian—and liberatory—dream of a World Wide Web of free information, it was also the beginning of the lucrative Internet. What made WCAG so influential was its unusual dual position: while the W3C positioned itself as an independent technical consortium, its web accessibility efforts gained substantial legitimacy from government support. The Web Accessibility Initiative, which would create WCAG, was chartered at the White House in 1997. When WCAG launched in 1999, US Vice President Al Gore's promise took precedence: “The U.S. Government intends to work closely with the World Wide Web Consortium to ensure that government information and services are accessible.” Despite its industry-driven, standards-focused appearance, WCAG's authority had government backing from the start—a connection that would prove crucial as WCAG became the foundation for accessibility regulations in the decades that followed.

One of the great paradoxes of US civil rights legislation like the ADA is that it is enforced through two primary avenues: administrative law and the judicial system, not through fines or other traditional levers for regulation. WCAG's regulatory adoption was no different: a version of WCAG was adapted into Section 508, requiring federal digital content to be accessible. As a result, compliance with WCAG became essential for federal and educational contracts. Employment and public education, covered by civil non-discrimination law, also required accessible software—at least on paper. Under the ADA, private companies face lawsuits when their sites break with screen readers, though enforcement typically targets organizations large enough to pay settlements. Through every enforcement channel, and even in the jurisprudence, WCAG became the de facto definition of digital accessibility.

All this regulation originates from guidelines published by an organization that steadfastly avoids presenting itself as regulatory. Despite government ties, the consortium cultivates an appearance of independence that aligns with web practitioners' persistent hostility to government intervention. Even those whose jobs depend on accessibility regulation hesitate to name the 'R' word: in this case “regulation.”

When Guidelines Meet Practice

In late 2024, when I spoke to most of my participants, web accessibility was coming off some of its best years: the COVID-19 pandemic forced much of daily life online while key legal decisions renewed industry interest. This progress created its own obstacles. Companies began selling accessibility overlays—automated widgets that float in the corners of websites, promising lawsuit protection by providing features like enlarged text and heightened contrast. Accessibility professionals and assistive technology users found these widgets failed to fix problems while managing to worsen navigation, serving as legal shields rather than actual accessibility improvements.

“I hate overlays,” said Melody, a Blind practitioner who works primarily as an auditor and a tester. “They're the worst type of tape you can find to very much not fix most of the problems,” she explained. “They're basically scams and, you know, marketing shenanigans that are like, 'Use our tool to get away with being fined and pay us instead.’” This frustration was universal among the Blind practitioners I interviewed—one software engineer noted overlays feature in a quarter of accessibility lawsuits. “They're paying to get sued,” she said bluntly. For the assistive technology community, these widgets represent compliance theater: creating the appearance of accessibility while actually making navigation harder for the very users they claim to help.

Overlays flourish as lawsuits become more common, exploiting companies' compliance fears. Yet practitioners rarely mention WCAG or regulations directly, despite these creating demand for their services. This reflects both the tech industry's anti-regulation stance and the guidelines' vagueness as practical blueprints for implementation. Instead, practitioners translate WCAG into design and development practices their colleagues accept, focusing on solutions rather than compliance language.

“The place of my work where WCAG guidelines are the most helpful is everything other than keyboard and screen reader interactions,” said Cameron*, an accessibility product manager at a large B2B company. Cameron found most guidelines too high-level—useful mainly as an enforcement mechanism when absolutely necessary. The guidelines establish a vague floor, they told me. Actual implementation requires deep knowledge of assistive technologies and the codebase in question.

WCAG also has significant gaps. Research shows it addresses almost exclusively the needs of blind users while neglecting other disabilities. The needs of users with cognitive disabilities face particular neglect, echoing decades of other accessibility gaps in architecture and design. While WCAG includes guidance for reversible errors, middle-school reading levels, and seizure prevention, these appear only at level AAA (the highest and most comprehensive tier) which most companies ignore.

Conformance calculations based on error counts can also distort priorities, inflating trivial issues while obscuring critical problems. Cameron encountered this firsthand. “I was handed [an audit] of 600 issues, none of which were ‘users cannot actually read or edit documents,’” they recalled. “It was everything else. It was, like, this toolbar has an unlabeled button and, like, this menu over here isn't quite doing the right thing and the color contrast and all this [other] stuff.”

Despite practitioners' deep knowledge of WCAG, they often find the guidelines themselves paradoxically both too prescriptive and too vague—useful as a baseline or enforcement tool but inadequate as implementation guidance. Web accessibility is no longer primarily a technical challenge but a question of resources, commitment, and execution. The technical solutions exist, but what's often missing is the institutional will to implement them.

Accessibility Under Threat

That missing institutional will would soon face an even more fundamental threat. All of this research occurred just as Trump was elected, and the new administration wasted no time in removing any references to accessibility on any federal site it could get its hands on. As more and more sites are built using non-WCAG-compliant site builders and LLM-based code tools, web accessibility is already on the decline. The growing backlash to disability rights is already changing the work of web accessibility. Practitioners find themselves not building for the future but fighting to preserve the meager achievements of the past.

In 2024, new regulation was added to the ADA to enshrine WCAG 2.2 as the accessibility standard for state governments, local governments, and public schools, making accessibility lawsuits easier to bring and simpler to win. Nonetheless, it would be impossible to say that the tech industry valued accessibility, or viewed it as anything other than a set of values that can be put on and removed when convenient.

“A lot of this work was often overlooked, or not considered even part of my career promotion,” said M, a software engineer who led accessibility efforts at their startup. M found it deeply discouraging that their accessibility work was dismissed as merely a hobby or passion project rather than legitimate professional work. “I was functionally doing two people's jobs. They only wanted to pay me for one, so I went back to doing one person's job.” Like other forms of maintenance work, accessibility is rarely taken seriously within organizations—a problem of resources and institutional commitment that no technical guidelines can address.

For most practitioners, 2025 has brought little but despair. After COVID-era restrictions were lifted by the federal government, many people with disabilities saw the care and excitement about their rights dissolve. The administration has begun treating accessibility as just another “woke” diversity initiative to dismantle, even coining the term “DEIA” to better target established protections. Despite the clear legal mandates of the ADA, federal agency procurement, the primary enforcement mechanism for web accessibility, now faces uncertain supervision.

It’s the end of that era, the era of government leadership on civil rights and inclusion, the era that made both the web and the W3C. Yet amid this political retreat, the W3C continues its work. Weekly meetings develop the next version of WCAG, with meeting minutes publicly available in keeping with the web's tradition of openness. Vanderheiden himself remains on the committee, continuing to shape web accessibility alongside new generations of practitioners. This persistence points to accessibility's unusual position: while government enforcement may falter, the standards endure—the web's architecture outlasting the politics that shaped it. The legacy of HTML built to be accessible from its earliest days continues, regardless of political shifts.

What I learned from the practitioners I spoke to is that accessibility work is hard, and often thankless, and easy to critique—but the vision of an accessible World Wide Web remains surprisingly persistent, even beyond the human-centered Internet. As more and more companies invest in autonomous agents for Internet navigation, I often wonder if future developers of these Internet navigation bots will lament the ways in which the web remains inaccessible to humans and machines alike—and how accessibility standards will continue to shape what is possible online. It’s clear to me that the guidelines will continue to evolve, even as enforcement changes, and that new and unexpected avenues will continue to arise. It's the end of an era in government-driven accessibility, but not the end of web accessibility itself.

---