I've been watching "unhosted" for a while, and the key thing to know is they're talking about data. The idea is that web apps following the "unhosted" pattern will get/put data from your storage. And this is an api that presumably storage providers/services would implement to enable that. So you'd either host a storage provider on your server somewhere, or pay a provider of such a service.

This is vastly different than some web app, ie a single page app, storing data on a cloud database somewhere (DBaaS for example). Because, the web app owner still owns and controls that cloud database. I'm very interested to see where this goes.

Edit: reading the docs, I have a lot of questions. This appears to be wrapper for browser local storage, that can then sync to a third party provider (local first). But browser storage is basically a key value store. Also reading the docs, they talk about using data modules defined with jsSchema.

That's all fine, but consider the use case for a note taking app. I store all the users data as json objects, that are evenutally persisted to the storage provider. So to query the users data on a new device, to get last weeks notes for example, I have to pull down all the user note data. So basically with this model you are pushing around large sets of json from device, with no way to query or get at a range of objects for example.

This is less appealing than it first seemed. Am I missing something?