Web Pages should be Linguistic Objects

[azaroth42]

As it's highly unlikely they contain no text whatsoever, even a title.

[azaroth42]

... but do we care about their language?

[azaroth42]

... yes because it's otherwise totally inconsistent with #111.

[workergnome]

Just to be difficult--what do we do with internationalized web pages, that can be in many languages?

[azaroth42]

At the moment we haven't actually given an example of language. Do we expect it to be singular or multiple? If it's multiple, I question how much use the information actually is, as you couldn't tell whether the page was mostly English with some Latin, or mostly Latin with a few English words here and there.

If the page handles language negotiation correctly, then there'll be a different URL for each language, and I would propose to follow IIIF/api#942 and reference each individually.

[azaroth42]

Re-closing, nothing that there is an ongoing issue that CRM is terrible at describing digital "things" that can change state, or have multiple states at the same time. There is the thing that is the home page of an object, and the linguistic content that is carried by that thing, in the same way that books can carry text. The current pattern avoids the complexity, but conflates issues that will likely need to be teased apart in the future.