Items

For us, Items represent learning resources in somewhat abstract way. They represent a specific set of ideas - without limiting to a specific artifact.

For example, the item "Sapiens" is a single item that is available in multiple artifacts/formats: a book you can buy, a podcast you can listen to, a video you can watch, a summary you can read etc. When I want to lookup my friend's review for Sapiens, I don't care which format they had consumed it in.

Any book catalog runs into this problem because they need to track various editions of the same book. OpenLibrary for example distinguishes between "Works" and "Editions".

We of course have to deal with many more formats or artifact types (31 as of now).

We avoided the naming problem for items by giving them a unique randomized UUID and storing the human-readable name. But the rest of the schema is not easy either.

First, we need to solve how best to represent links because URLs are not good enough.

The overall topic schema becomes:

id
name
description
image
links
topics
creators
difficulty
rating
tags
reviews