Even after reading several of the github linked specs and examples, I still struggle to understand what this intends to be, how it will work with e.g. CORS, or

Even after reading several of the github linked specs and examples, I still struggle to understand what this intends to be, how it will work with e.g. CORS, or how anything will be able to flexibly work with other things playing in the same data-sandbox. Like, we already have XML DTDs and schema.org and whatnot. Everything appears to be in place to build everything that I can see this doing, and has been for decades. But it hasn't turned into anything because the problem isn't how to store the data, it's how to use it, and allowing everyone to manipulate all data seems to repeatedly be demonstrated as a failure. Is it really just "let websites manipulate my data, stored on my machines"? What would possibly incentivize websites to do so, rather than pull in data and subtly break it for others (intentionally or by accident)? Users are absolutely not going to understand why site X broke site Y, only that site Y is broken. Or am I missing something fundamental? Entirely possible, I can't figure much out at all.