The document is the natural representation of data, and relational databases have been dominated by rows and columns since the 1970s as a way to optimize data access. However, technology has moved on, and systems are now capable of managing documents in JSON format, allowing for efficient storage, querying, and manipulation. Document databases like MongoDB can hold all the information together in a single document, making it easier to retrieve and manipulate data without the need for multiple lookups or joins. This approach also aligns well with modern software development practices that treat abstract things as objects, where documents are fundamentally a collection of data that can be easily retrieved and manipulated. By using documents, developers can avoid the overhead of ORMs and achieve better performance, making it easier to iterate on their application design.