mattgarrish
commented
1 day ago For the EPUB metadata, this also requires additional logic to parse and split the value provided in dcterms:conformsTo which while not difficult, requires additional processing that seems unnecessary.
Looking at the original issue, I believe this is related to why the strings are being parsed rather than matched. It sounds like kobo would reject a string that isn't a full conformance statement, but this processing could match invalid conformance statements and even statements that don't have anything to do with accessibility or that use WCAG-like terms for an internal standard (e.g., there's no reason "Level A" couldn't be part of a conformance statement that has nothing to do with accessibility). The techniques assume there's only an accessibility conformance statement and that it's a valid one.
It might be safer to pattern match off the full conformance string/url since the epub and wcag versions, and the wcag levels, are always in predictable spots.
On #471 @wareid made several comments.
Could we please get clarification?
First: • For the EPUB metadata, this also requires additional logic to parse and split the value provided in dcterms:conformsTo which while not difficult, requires additional processing that seems unnecessary. Please elaborate on what is not necessary. Second: Mentions of metadata properties in the ONIX and EPUB metadata docs are not made in the main document, which adds to the confusion. We have a note that not all of the Additional Information is catalogued in the guidelines. Listing all of them would make the guidelines much longer and many of these are not frequently used. What do you suggest? Add or remove reactions