If Accessibility was treated like engineering or a right, not a favour!

If accessibility were treated like engineering, we would not still be writing newsletters in 2026 saying:
“Here are the five most common accessibility issues. Please try harder.”
We have known these issues for decades.
Bad contrast.
Missing alt text.
Unlabelled form fields.
Anonymous buttons.
Broken document structure.
Bad ARIA.
These are not mysterious edge cases. They are not advanced accessibility theory. They are not the sacred mysteries of inclusive design.
They are recurring, measurable, preventable failures.
And yet year after year, we keep finding the same things, publishing the same checklists, giving the same talks, writing the same advice, and then acting surprised when the same errors come back again.
At some point, this stops being an awareness problem.
It becomes an engineering failure.
If a build pipeline can stop code shipping because of a syntax error, it can stop code shipping because a form field has no label.
If a CMS can warn you about SEO, image size, broken links, cookies, analytics, metadata, and whether your headline is “engaging” enough, it can stop you publishing an image with no useful alternative text.
If a design system can enforce brand colours, spacing, typography, and component behaviour, it can enforce contrast, names, labels, headings, and actual assistive technology output.
This stuff should not depend on whether the right person remembered the right checklist on the right Tuesday afternoon.
That is not engineering.
That is hope.
And accessibility is still treated far too often as hope.
Hope the developer cares.
Hope the designer remembers.
Hope the content editor has been trained.
Hope the procurement team asks the right question.
Hope the disabled user complains politely enough.
Hope the organisation feels embarrassed for long enough to fix it.
But disabled people’s access to information, services, money, medicine, education, work, travel, culture, and public life should not depend on hope.
It should not depend on kindness either.
Kindness is lovely. I am very pro-kindness. But accessibility is not a favour. It is not a nice extra. It is not a charitable flourish added after the real work is done.
It is part of whether the thing works.
A button without a name does not work.
A form field without a label does not work.
A page that cannot be navigated non-visually does not work.
A service that silently excludes people and then calls itself digital transformation does not work.
The accessibility field has spent a generation educating the people who were willing to listen. That work matters. But it has also allowed education to be mistaken for remediation.
A checklist is not a control.
A webinar is not a fix.
A guideline is not an interlock.
A request for vigilance is not a safety system.
If accessibility were treated like engineering, the known mechanical failures would be blocked at the point they are introduced. The tools would carry the discipline. The frameworks, CMSs, design systems, site builders, component libraries, and publishing platforms would stop accepting defects as normal output.
And then human accessibility expertise could move where it actually belongs up a level.
Not endlessly re-explaining alt text and labels but asking the real questions.
Can the journey be completed?
Can the user recover from error?
Does the service make sense through assistive technology?
Does the system preserve dignity?
Where do people give up?
What information casualties are we not counting because the person simply closed the tab?
That is the scandal at the heart of this.
The casualties are real, but invisible. There is no wreckage. No incident number. No manifest. No investigation. Someone cannot complete the task the system was built to provide, and the system records nothing.
Accessibility treated as a favour says: “Please care more.”
Accessibility treated as a right says: “You do not get to exclude people by accident and call the system successful.”
Accessibility treated as engineering says: “Known failures should not be shippable.”
That is the change I want.
Not more guilt.
Not more awareness theatre.
Not another annual ritual of telling the same conscientious people what they already know.
Build the horn.
Install the interlock.
Count the casualties.
Stop pretending advice is the same as repair.
Because if the same five failures are still dominating the web after twenty-seven years, the problem is not that disabled people have failed to explain themselves clearly enough.
The problem is that the system was never forced to learn.