Orchestra / The future of UI and accessibility
More ways
to finish.
Interfaces have changed what people need to learn. Agents could change what they need to operate. These six articles ask who benefits, what remains difficult and how to tell whether the work actually got easier.
Start with the argument →- 01Read →
History and human capability
Is UI holding us back?
New interfaces can remove work or move it into checking and correction. A historical argument for judging the whole task.
Historical synthesis - 02Read →
Accessibility today
Web accessibility: can people finish the job?
An accessible button is useful. An accessible way to finish is the point. What the current evidence tells us about agents and access.
Research and proposed evaluation - 03Read →
WebMCP and shared state
Stop making agents hunt for buttons
WebMCP lets a page expose actions to a browser agent. Radar shows how that can keep the person and agent working in the same place.
Radar implementation and draft specification - 04Read →
Commander / First Officer
A spoken promise is not a completed action
A Cargo recovery request becomes an action, a checked result and a spoken report. What this bounded demonstration can teach app designers.
Recorded sandbox demonstration - 05Read →
Voice through MCP
Say the job once. Keep control of the result.
Dictating to an assistant can become a way to direct connected tools. Our proposed experiment starts with a scoped read and a draft.
Proposed experiment - 06Read →
The future of UI
Your work should survive a change of interface
Speak, inspect, correct and continue. A future interface should preserve the task when the person changes how they interact.
Design position and research agenda
Living reference
Different routes. Different responsibilities.
These approaches coexist. Voice is an input method; WebMCP and MCP provide different tool connections. This is a map of responsibilities, not a chronology or performance ranking.
- Event-driven controls
- A person chooses controls; events trigger app actions. The person directs the sequence.
- API / CLI
- A person or program specifies operations. Suitable for repeatable work when the operations can be defined.
- WebMCP
- A compatible browser agent calls actions exposed by a web page. The page can remain a shared place to inspect the work.
- Voice over an app
- A spoken request reaches an app-specific integration. Commander supplies a bounded example.
- Voice through MCP
- An assistant could turn speech into authorised calls across services. This is the next experiment proposed in the series.
How we will judge progress
Count the effort to finish.
Compare the same task through the existing interface and the proposed route. Define completion first. Record assistance, errors, corrections and the effort of checking the result. Account for familiarity, input method, cost and connectivity. Include people who use the access methods being claimed.
The proposed experiments are a scoped Commander read through MCP, correction after changing input, and a draft that crosses app boundaries. These experiments and a participant study have not been completed for this series. The current evidence is historical research, the reported WebAIM scan, inspected implementations and an existing Cargo recording.
Yann VR is the accountable editor. Orchestra builds the demonstrated systems; they are not independent evaluations or client outcomes. Primary sources appear beside claims. Source notes are our summaries. Conceptual figures carry no measured performance ranking.
Evidence reviewed 5 October 2026. Material source changes, reproducibility failures and substantiated reader corrections trigger a review of the text and figures. Previous observations retain their dates. First published 6 October 2026; no revision since.
Send a correction or discuss a task →