@Felthry @myst It’s not so hard. Consider a database of User entities. Each User is a table. the rows are keys (userID, avatar, accountIsPrivate) and the columns are values associated with the keys (0123, “imageurl.png”, true). Sometimes the keys point to lists, or to IDs identifying other parts of the database. For example, a Comment entity might be a internal numeric ID and its text contents. The value of the comments key of a User might be a list of comment ID numbers.

@Felthry @kellerfuchs So if you have, say, a webapp, you're going to need to handle things like, say, people trying to do two conflicting operations at the same time - like making accounts with the same username. A way to handle this is talking to a networked database, like, in this case, PostgreSQL. It can handle things like what operations actually get to happen and ensure eventual consistency. Postgre's a relational database because the data is stored in tables with relationships to each other - the main alternative, I believe, is things like NoSQL with servers like MongoDB.