Use case: Use semantic data for accessibility

[opyh]

As discussed in https://github.com/Maps4HTML/MapML/issues/130 and in https://discourse.wicg.io/t/physical-accessibility-data-in-maps/4852/5, here is a use case for map + screen reader accessibility:

• Have a map with 15 places of interest
• The places of interest are features and have very basic semantic RDFa markup
• The features are typed as subtypes of schema.org Place (e.g. Hotel, FoodEstablishment)
• The prototype adds an aria-label on the map element that summarizes the visible information
• An exemplary aria-label would read: "a map displaying two hotels, one shopping center, and thirteen more places"

MapML/RDFa feature code example:

<feature vocab="https://schema.org" typeof="Hotel">
     <properties>
        <span property="name">An awesome hotel 🏨</span>
     </properties>
     <geometry> ... </geometry>
</feature>

We could later add use cases related to physical accessibility, e.g. letting the user agent change the visual appearance of displayed features depending on their physical accessibility features, as outlined in the presentation.

[prushforth]

I'd like to get comments from others about this use case, but my initial impression is that it would fit as a "content author use case".

Aside: although the distinction can be blurry, I've often though of "map author" as a role that could (sometimes) be distinct from "html author", especially if the content is separate from the page into which it's being included, as foreseen by the <layer src="..."></layer> construct.

Maybe there's an assumption here that a

<feature vocab="https://schema.org" typeof="Hotel">
     <properties>
        <span property="name">An awesome hotel 🏨</span>
     </properties>
     <geometry> ... </geometry>
</feature>

element is a requirement, so please advise on that topic as well, and how it may fit into the abstract use cases presented by our document. Maybe there needs to be a use case like "Need a semantic <map> element to be the container for maps", or something like that. That way we have a way to discuss how to get from here to MapML, <layer>, <feature> and so on.

[Malvoz]

@opyh Just to clarify, this isn't about requiring user agents to display accessibility features based on semantic data is it? But rather to "simply" support the use case of marking up features using semantic data?

I think we can narrow this issue down to the use case of supporting semantic data (RDFa, Microdata, Personalization attributes?, etc.) in general, as that'd cover the described use case (semantic data for accessibility).

• The prototype adds an aria-label on the map element that summarizes the visible information
• An exemplary aria-label would read: "a map displaying two hotels, one shopping center, and thirteen more places"

If those are suggestions for the prototype, please raise an issue in https://github.com/Maps4HTML/Web-Map-Custom-Element (see related PR: https://github.com/Maps4HTML/Web-Map-Custom-Element/pull/151) and we can discuss!

@prushforth

Maybe there needs to be a use case like "Need a semantic <map> element to be the container for maps"

Or rather, we can add a note about potentially requiring an implicit ARIA role in the Conclusion for Capability: Embed an interactive map viewer, using HTML markup (and label it "Accessibility: potential improvement").

[opyh]

Yes, the idea was to add RDFa support in general, because it would allow for a lot of interesting things to happen. That it makes a lot of map accessibility use cases possible (e.g. browser plugins that change maps on the fly) is a subset.

From what I see, aria-label etc. isn't part of the spec yet, either (it should be though).

Should we split this into two use cases then? One for RDFa, one for direct accessibility label support?

[Malvoz]

From what I see, aria-label etc. isn't part of the spec yet

The MapML spec does allow global aria-* attributes on the <map> element.