With the ability to generate Guided Navigation Documents from the HTML resources of a publication, we're also in a position where we can infer a list of languages.
This can be based on:
- the languages that we encounter while parsing HTML
- ordered by the number of strings (no need to work with characters IMO)
Since this is potentially riskier than the rest, we should never merge this metadata with the rest and only use it when inferred metadata are split.
For this purpose, we can use language with a list of ordered BCP-47 codes under https://readium.org/webpub-manifest#inferredAccessibility.
With the ability to generate Guided Navigation Documents from the HTML resources of a publication, we're also in a position where we can infer a list of languages.
This can be based on:
Since this is potentially riskier than the rest, we should never merge this metadata with the rest and only use it when inferred metadata are split.
For this purpose, we can use
languagewith a list of ordered BCP-47 codes underhttps://readium.org/webpub-manifest#inferredAccessibility.