The discussion revolves around an accessibility issue with heading labels in Discourse, specifically affecting screen reader users. adress reports that while VoiceOver can navigate to reply headings, the heading text does not clearly identify the purpose or content of each reply, making it difficult for users to determine the relevance of each reply without visiting it. This issue was previously reported in another topic. chapoi asks for clarification on the expected behavior, and adress explains that users should be able to know which post is being replied to. chapoi shares their experience with VoiceOver, and adress emphasizes the need for blind users to distinguish which reply button is associated with which post. adress mentions a related issue, and awesomerobot asks if the issue is similar to another previously discussed problem. adress confirms that the issue is with screen readers not being able to distinguish between links, buttons, and elements with the same repeated name.
Issue: While VoiceOver can navigate to reply headings using quick navigation keys, the heading text does not clearly identify the purpose or content of each reply.
Specific Behavior:
● Headings announce profile names but lack context about reply content
● Users can find replies but cannot efficiently identify which reply contains relevant information
Why High: Makes discussion reading very difficult and confusing as users must visit each reply to determine relevance.
But how? What do you expect to hear? I’m not familiar at all with screenreaders, hence I’m asking.
When I turn on voiceover in safari, and Im on the reply button element, it says something like:
For example, sighted folks would know which of the reply buttons to activate because of visual grouping. We need a way for a blind users (using a screen reader) to know for certain which reply button is associated with which post. Right now, people have to listen to them all.
@adress is this using an element list style of navigation similar to what was mentioned in the other issue? in this case, the landmarks are difficult to distinguish from each other… here’s an example from VoiceOver