OpenMRS uses Unit testing. Why is it important and why should we learn it?

By



Before talking about unit testing and why it is preferred by large and open-source organizations like OpenMRS, let me introduce "code testing" to those without a background in code development. 
Much like field testing a vehicle before launching it to the consumers, or testing a vaccine before releasing it to the public, a code needs to be tested thoroughly before it is deployed and shared with clients. How thorough? As thorough as it can be. It is very important to test every possible case and every code block. All accepted input types should be tested and their outputs respectively. It is crucial for the longevity of the code and the developer producing it, to test their code before releasing it. It prevents any future recalls and resource allocation for fixing something that should have been caught in a testing phase. 




In the old days of programming, manual testing was one of the only options available. Even though it is still used by many, it is not wise to manually test your code especially if the code is part of a larger organization or code structure. Manual testing is tiresome, inefficient, and not thorough. Many automated testing methods are in use today and unit testing is one of them. In my research of my foss project, I noticed that OpenMRS heavily relies on unit testing and is the core component of their testing methods. 

In order to deploy unit testing, you need to first have a set code structure that allows unit testing to be efficient. Small modules of codes are the best practice for an efficient unit testing follow up. These small modules perform one specific task each for unit tests. A unit test tests a small module by encapsulation and does not require the module to connect to any external resources like a database or other module integration. Any external dependencies, meaning any module or component that relies on another component to work, need to go through a "mock" process. They need to be faked in order to unit test a module. A fake implementation of the dependency is used to allow the component to be tested, run properly, and in turn, tested properly. 

A quick example of separate components that could be unit tested in a code is:

public class CarInventory {
    public Car Tesla(string model) {
        var car = new Car
        {
            Make = "Tesla";
            Model =  model;
        };
        return car;
    }

public Car Ford(string model) {
        var car = new Car
        {
            Make = "Ford";
            Model =  model;
        };
        return car;
    }

}

In this example, Tesla() and Ford() are two separate modules. They can individually be tested by unit tests. This allows simple readability and execution of tests. 



Comments

Post a Comment

Popular posts from this blog

Make way for GPLV3

By
  GPLv3 is the new GPLv2 A little more than a decade ago, when GPLv3 was first introduced, a debate sparked between the fans of GPLv2 and GPLv3. Linus Torvald, the inventor of Linux, sided with GPLv2 and wrote several emails and threads on why his choice is justified. Richard Stallman, the founder of the Free Software Foundation, chose GPLv3 as his license of choice between the two. The main reason why Torvald didn't approve of the new and improved version of the GPL license was the overreaching aspects it contained. GPLv2 allowed hardware manufacturers to embed a signature key in their chips which would be needed to run free software on a machine. This specific signature key however limited the flexibility of the developer community since they couldn't update or modify the code and run it on the same machine. This limitation set forth by hardware manufacturers was likened by Linus and his followers. GPLv3 tackled this constraint by including an anti-tivoization clause in the l...
Read more

When a code rot can be the difference between life and death.

By
Image
 In 2018, a 9-year-old girl who had type-1 diabetes her entire life was being misled by her trusty medical applications on her smartphone. If these misleadings were not investigated by her father, Jeremy Vaughn,  she could have faced critical and fatal medical consequences.  Her medical check-up app with a continuous glucose monitor (CGM) failed.  “Features were disappearing, critical alerts weren’t working, and notifications just stopped,” he stated. As a result, his nine-year-old daughter, who relied and lived her whole life on the CGM alerts, had to live on instincts.  These apps were downloaded by Jeremy in 2016 for his daughter but by 2018, despite regular updates, these apps were now outdated and almost non-functioning. Jeremy checked the reviews on these apps and several patients had complained about malfunctioning apps and how they directly affected their health in a critical way. All this was due to the rotting code in these applications. They were regu...
Read more

Privacy and Personal Data

By
Image
  Data privacy or information privacy is a branch of data security concerned with the proper handling of data – consent, notice, and regulatory obligations. More specifically, practical data privacy concerns often revolve around how the data is being shared with third parties and the legality of its collection and storage. The world runs on transactions. Buying and selling. What's more important than introducing your product in a new market to try and gather customers? Introducing your product in a new market with guaranteed customers. And how are these customers guaranteed? Big Data.  When you shop at Amazon, do you ever see a "customers like you also bought ...."? The item being introduced to you have a higher chance of being sold than to a random customer. It's very simple in how it works, but extremely effective. Look at it this way: A company "X" does not have any information on its customers and their interests. In contrast, a company "Y" has...
Read more