ERROR RSC-005 thrown on role="doc-epigraph"

[atroia]

Used role="doc-epigraph" in a section tag and I'm getting **ERROR RSC-005 : Error while parsing file: value of attribute "role" is invalid; **. I'm not seeing that this ARIA role has been deprecated or done away with.

[atroia]

role="doc-cover" also throwing the same error.

[rdeltour]

doc-epigraph seems to be an omission on the roles allowed in section as defined in ARIA in HTML, wdyt @mattgarrish? If yes we should file an issue there.

doc-cover is an image role and MUST NOT be set to a section, so that’s a legit error.

For reference: the rules of where DPUB ARIA roles are allowed in HTML are defined in the ARIA in HTML specification.

[TzviyaSiegman]

Good catch @rdeltour. @mattgarrish will you please log an issue on ARIA in HTML and the validator?

[mattgarrish]

Only roles that have inherit from the abstract landmark role are allowed on section, so the error is correct (see the superclass roles in the DPUB-ARIA definition tables).

doc-epigraph inherits from the abstract section role, so it falls into the document structure category. It can be used on blockquote, div, etc.

(The section superclass is not the same as the section element, in one of ARIA's odder quirks.)

[rdeltour]

Only roles that have inherit from the abstract landmark role are allowed on section

Then there are inconsistencies in ARIA in HTML: e.g. doc-abstract, doc-example, doc-qna, and others inherit from the "section" role and are allowed on the section element…

[rdeltour]

Only roles that have inherit from the abstract landmark role are allowed on section

I'm not sure I understand why that should be the case however. Why the section element would be limited to landmark-inheriting roles? This isn't the case for other non-DPUB ARIA roles (see alert for instance).

[rdeltour]

btw @atroia, while we're discussing all this (😊), you can always use the role doc-epigraph on the blockquote element, which likely makes the most sense in any case.

[mattgarrish]

Why the section element would be limited to landmark-inheriting roles?

When we created the vocabulary, we were advised to assign landmark roles to the major sectional semantics and assign section to items that are not. That's why you won't find any mixing of section with the non-landmark roles in the examples.

I'm not sure why the ARIA document has mixed them, but we also weren't asked about the mapping in that document (that I remember). Definitely something we should clear up, as there may be assumed conditions on the use of the non-landmark roles (like no headings).

[atroia]

@rdeltour Thank you! That actually works out perfect, since I have an opening blockquote tag right below the section tag that I thought I needed. @mattgarrish I had no idea about the section superclass vs element, I still have so much to learn about ARIA.

[LaZay]

In HTML, a section can be recursively nested into a section. It is not possible to postulate that a section is ALWAYS used as a landmark. The parallel between section and TOC was just the most powerful and pedagogical example to show the interest of new HTML5 sections. It was not a requirement though. HTML5 has never put such a semantic constraint on section element definition.
https://www.w3.org/TR/2011/WD-html5-20110525/sections.html#the-section-element

[flittle8]

Hi, wondering if you can confirm if it's legitimate to add epigraph semantics to the section tag as per the DAISY knowledge base?

[mattgarrish]

Yes, it is allowed on section.

The ARIA in HTML document is the authoritative reference for which roles are allowed on which elements. We've worked out the discrepancies that used to exist.