The Ledger Still Open: Assistive Technology at HCI 2009 and the Audit We Have Not Closed

In September 2009, delegates travelled to Churchill College, Cambridge, for the twenty-third annual conference of the BCS Human–Computer Interaction group, published in proceedings as People and Computers XXIII. The programme was wide: mobile interaction, tangible computing, affect and evaluation. Tucked into the timetable was a smaller strand on assistive technology and inclusive design — sessions with modest audiences and, judging by the subsequent citation record, even more modest afterlives.

Seventeen years on, that strand reads less like a historical curiosity and more like an unfinished audit.

Proceedings as a record of what was already known

"We did not know then" is the standard defence offered whenever an accessibility failure surfaces in public. The archive makes that defence difficult to sustain. The 2009 assistive-technology papers set out, in ordinary and often unglamorous detail, the failure modes that UK practitioners still meet: interfaces that presume a single mode of input, documentation that treats compliance as an arrival rather than a departure, and evaluation studies whose participants are already comfortable with the technology under test.

None of this was heretical. It was not even particularly contested. It was simply filed away, in the way that conference proceedings are always filed away, and the field moved on to the next attractive problem.

Three warnings that kept recurring

The strand circled the same concerns with unusual consistency.

• The compliance trap. Several papers argued that a service could satisfy a checklist in full and remain unusable in practice — that standards describe a floor, not a ceiling, and that organisations routinely mistake the one for the other.
• The evaluation gap. Samples were small, recruitment was frequently a matter of convenience, and participants with lived experience of assistive technology were seldom involved in shaping the study itself.
• The maintenance problem. Accessibility was framed as a one-off retrofit rather than a property to be sustained; the papers anticipated that reorganisations, content churn and platform migrations would quietly erode whatever had been achieved.

A reader in 2026 will recognise all three. What is striking is not their prescience but their ordinariness. These were not predictions requiring unusual foresight. They were descriptions of observable behaviour.

The regulations arrived; the gaps did not close

The UK's legal position has been well documented elsewhere on this archive. The Equality Act 2010 established a duty to make reasonable adjustments; the Public Sector Bodies (Websites and Mobile Applications) Accessibility Regulations 2018 added specific obligations, monitored centrally and evidenced through published accessibility statements.

Those instruments matter, and they have changed practice. But they leave gaps that the 2009 authors would have recognised immediately. Enforcement rests heavily on complaint, which places the burden on the people least equipped to carry it. Accessibility statements are frequently stale, sometimes inaccurate, and rarely written for the reader who needs them most. Third-party embeds, legacy PDFs and inherited intranets sit awkwardly at the edge of any monitoring regime. Small authorities, whose digital estates are large and whose budgets are not, publish statements they have neither the staff nor the tools to substantiate.

The commercial sector remains governed largely by the general duty rather than by any specific accessibility standard, and the European Accessibility Act — whose obligations began to bite for products and services placed on the EU market from mid-2025 — reaches UK suppliers only insofar as they export. Domestic consumers get whatever the market and the courts provide.

Where 2026 has moved the ground

Three developments have changed the terrain since the 2018 Regulations were drafted. The first is the arrival of generative and conversational interfaces, whose accessibility properties are poorly understood and unevenly documented; a chat window with no accessible name is a familiar sight on otherwise well-regarded services. The second is the commercial overlay industry, which reproduces an old instinct — that a technical layer can substitute for design judgement — in a new form. The third is procurement pressure: as councils and NHS trusts consolidate platforms, decisions about accessibility are increasingly made by suppliers, far from the users who will live with the consequences.

Each of these would have been legible to a 2009 delegate. The overlay debate, in particular, is a rerun of an argument conducted in the corridors of that Cambridge venue, with the same assertions and the same missing evidence.

Why historians should not close the file

There is a tendency to treat conference proceedings as a record of what a field once thought, of purely antiquarian interest. The HCI 2009 archive supports a different use. Because the papers specify problems rather than merely naming them, they function as a baseline against which later practice can be measured. That makes them useful to technology historians, who can trace which warnings were absorbed and which were discarded, and useful to UX leads, who can test whether their own organisation has quietly re-created a documented failure.

Reading the strand in this way is not an exercise in retrospective scolding. It is an exercise in accountability, which is a narrower and more practical thing.

A modest re-audit

A proposal, offered in the spirit of the archive. Take three papers from the 2009 assistive-technology strand. Apply their central claims to a current UK public-sector service — a council tax portal, a patient-facing NHS tool, a university admissions system. Publish the results, including the parts that reflect badly on present practice.

That would be a genuinely retrospective exercise: not a celebration of a conference, but a continuation of a conversation that was never properly concluded. The debt described in these proceedings was not incurred in 2009. It was merely documented there. The ledger remains open, and the interest has been accruing for seventeen years.