Webdev

The bookmark itself is tied to a URL, and anything else related to the bookmark, such as the title, the scraped content (if the service scrapes on your behalf), highlights and annotations are stored as additional metadata linked to that URL.

There are some unfortunate restrictions that come with this data model.

Let’s take the example of comments.

Comments by their nature are distributed; the same article can be shared on any number of websites for any number of different users and communities to discuss. Especially in the case of tightly focused communities, the commentary on an article is often just as valuable as the article itself.

When the data model is anchored around the URL, highlights become tightly coupled to that URL, and this tight coupling ignores the distributed reality of comments, leaving them with no real place to exist in the data model.

Ideally, highlights made on an article and comments saved about an article should be easy to connect and view together, because it is the article itself, and not the URL, that is the common denominator (the URL is an imperfect proxy for the article).