Discuss the main categories of data models. What are the basic differences among the relational model, the object model, and the XML model?

The main categories of data models are $\textbf{high-level}$ or $\textbf{conceptual data models}$, $\textbf{low-level}$ or $\textbf{physical data models}$ and $\textbf{representational}$ or $\textbf{implementation data models}$. Another class of data models are $\textbf{self-describing data models}$.

$\bullet$ Conceptual data models provide concepts understandable to many users, such as entities, attributes and relationships. Most popular high-level conceptual data model is $\textit{entity-relationship model}$.

$\bullet$ Physical data models describe details of how data is stored on the computer storage media such as magnetic disks. These models are meant for computer specialists which can better understand these concepts than end user.

$\bullet$ Representational data models may be easy to understand by end users, but they are still not too far away from the way data is organized in computer storage and that's why we can say they are between two extremes - conceptual and physical data model.

$\bullet$ Self-describing data models combines the description of the data with the data values. These models include $\textit{XML}$ and many of the $\textit{key-value stores}$ and $\textit{NOSQL systems}$.

$\textbf{Relational model}$ is representational data model used in traditional commercial DBMS. Data is represented by record structures ($\textit{record-based data model}$). $\textbf{Object-data model}$ is a new family of higher-level implementation data models closer to conceptual data models. $\textbf{XML model}$ belongs to self-describing data models. Therefore, the main difference of these three models is that each of them belongs to different category of data models.

The main categories of data models are: $\textbf{high-level}$, $\textbf{representational model}$ and $\textbf{low-level}$. $\textit{High-level}$ or also called $\textit{conceptual data models}$ use concepts such as enteties, their attributes and relations between them. Those are data that user can see and understand. Second one, $\textit{representational}$ or $\textit{implementation}$ (not a standard term) data models hide many details of data storage on disk but can be implemented on a computer system directly. So user can see / understand data from this model, but not all of them. And on the end, there are $\textit{low-level}$ or $\textit{physical data models}$ that provide concepts that describe details of how data is stored on the computer storage. Those concepts are meant for computer specialists, not for users.

The basic differences among the relational model, the object model, and the XML model are that in relational model data and relation data are shown in tables, the object model is inspirated by an object oriented programming so there are objects and there are classes. XML model is tree-structured data model.