Why do DTOs lead to temporal coupling?

In the comments section of one my recent articles I have been asked the following question: “Why do DTOs lead to temporal coupling?”. Seriously, why? Have we not been using them for years in Java without critically thinking about what could be wrong with them?

Many very similar houses placed close to each other
DTOs in practice. © writingfordesigners.com

As always, let’s start with a short piece of code that demonstrates the topic of discussion.

    StudentDao studentDao = new StudentDaoImpl();
    //update student
    Student student = studentDao.getAllStudents().get(0);


This block of code shows a very typical way to use DTO in the context of DAO operations. At first sight, the ability to fetch data in a form of an object, update it, and then send it back seems lucrative. However, let’s look at this code again from the perspective of OOP and Elegant Objects.

In the presented context, it is visible that a certain controller acts upon the Student DTO in various ways. First, it changes the name of the student by using a setter. Next, it sends the updated data to the database, by using a DAO instance, which extracts the data fields from the Student DTO. What exactly is wrong here and why DTO is the one to blame?

  1. First of all, being nothing, but a plain data container, the Student DTO is a dependent object. Being able to provide data, but not functionality or behavior (a combination of data and context) it loses independence as an entity. As a result, the Student DTO, just like any other DTO (read-only or not), requires context before it can be used. The object which provides context is conventionally called controller.

  2. Requiring someone to act upon it, due to its surrogate nature, DTO becomes a utility component of a controller, the necessity of which is created by the DTO itself. Consequently, the controller becomes an object that collects various DTOs, works with them however it needs to and then throws them away.

  3. The consequences of the procedural nature of the controller bring us back to the beginning of the article (The connection between DTOs and temporal coupling). The more DTOs the controller works with, the more complex it becomes and the more lines of code it ends up having.

By working with multiple data containers in one place, a controller becomes a procedural block of code that is intolerant to any changes, leading to increased temporal coupling and complexity (due to the increased context, introduced by different DTOs) and lower maintainability as a consequence.