Use git, use eslint, write unit tests.
Want to know the reasoning behind each tool? Keep reading!
Whether you are starting out as a developer or already have some experience, improving your coding skills is an ongoing endeavour. But what does it mean to be a better coder?
While there are many ways to solve a problem with code, some of them are better than others. As Robert Martin wrote:
Even bad code can function. But if code isn’t clean, it can bring a development organization to its knees.
Good code doesn’t just work. It is also easy to maintain and reuse. That means that others (or the future you) can understand this code easily, which in turn allows them to solve bugs, add features and refactor it faster and without degrading the whole application.
As such, code quality has a big impact on the happiness of everyone involved with the project: engineers, managers, product people and even endusers.
Being a better coder means writing better code, and getting there requires a lot of practice.
If you work in a well operated team, this practice usually comes through design and code reviews, but this is not always the case – especially if you are working on small pet projects which are not developed in a team.
If that’s the situation, you can use readily available tools that will help you experiment, get feedback and learn better coding practices.
These days, it’s hard to find software development teams that don’t use some tool for version management of their code. Git is one such tool and is probably the most popular. CVS’s track changes in the code and allows multiple developers to work in parallel and collaborate on the same codebase.
But what if I’m working alone? what if I’m just playing with some code to learn a new library or try out some ideas? I don’t need no git, right?
Small children learn by experimentation. They try out different actions in a safe environment, see the results and then try again something different.
The same approach can be applied to coding, and git is the tool that enables that.
When used correctly, git enables one to explore different solutions to a problem, try out different libraries, and experiment freely, with the knowledge that at any point in time it is trivial to roll back the changes, get back a good version of the code and start over if needed.
It’s like having an ‘undo’ button for entire change sets and is really liberating.
I recommend to use git for every project — even if it stays local and not uploaded to a remote repository like github.
If you already know git it will be trivial and if not you’ll learn a new and important tool.
Linting tools, like ESLint perform static analysis of the code and recommend on changes that can improve it according to a variety of rules.
They are used in software projects to maintain a consistent standard of code style and quality in the project.
However, with the right approach they can also help improving one’s coding skills even in personal projects.
I would divide the eslint rules to two major categories: code style and best practices. When working in a team it is very important to maintain a standard code style because it makes reading the code easier, but there is no one right code style and this category of rules is less important in the context of this article (I don’t think you’ll be a better coder if you use tabs instead of spaces).
The second category, however is very relevant for our subject. It include rules that are meant to prevent potential bugs, make the code more readable and thus more maintainable.
This is important for development, but if you pay attention these rules can also teach you what patterns to avoid in the future which is a great win. It’s almost like having an experienced developer reviewing every line of your code in real time and giving focused, well explained recommendations for improving your code.
Starting out with eslint can be a bit frustrating because there will be many warnings and errors and the reasoning behind each rule is not always clear, but remember that these rules were created by highly experienced developers and were refined over years of usage.
Note that the list of rules that are enabled can be customised, so I recommend to start with a strict set of rules and disable a rule only after you understand why it was created and what is the reasoning behind it.
As a side note I’ll add that even the rules in the “Stylistic Issues” section, which are said to be “quite subjective”, can make your code better (if not prettier).
Take for example, the rule max-lines-per-function which limits the number of lines that a function can have. Even though it is considered a styling rule, it has importance beyond just style. This is because a function having too many lines usually indicates that it has too much logic or too many responsibilities and should be broken into smaller, simpler pieces of logic.
Also note that there are plugins for ESLint which add rules that enforce best practices for more special cases like specific frameworks, functional programming, arrays, promises, etc. and it’s a good idea to get to know them too.
Unit tests have a major part in assuring the quality of code in software projects. They help avoid bugs and make refactoring easier.
But besides that, unit tests can also teach you to write better code. The reason for that is that bad code is usually also hard to test, so as you gain experience writing tests, you will always think about the testability of your code even if you don’t follow a strict Test-Driven-Development workflow. Let’s see two examples for how good testability goes hand-in-hand with good coding practices.
The first example is global variables.
Using global variables is considered a bad practice for various reasons which are not necessarily related to tests. Mainly, they cause the code to be less predictable, less manageable and less reusable as they create implicit, unregulated couplings between different areas of the code base.
In the context of unit tests, source code that uses global variables is more difficult to test. The reason is that to be effective, each separate test case has to run in a controlled, well defined setup to consistently produce the expected results. Having global variables around makes it tricky to control the test setup completely as they create implicit dependencies which are hard to control.
Because of this difficulty, when you keep unit tests in mind, you will tend to avoid using global variables and get better code.
For the second example, let’s consider a function with complicated logic and many arguments. Writing good tests for such function is going to be a nightmare, as in order to be effective we must check all execution paths in the function which require as many combinations of input values as possible. Breaking such function down into smaller, simpler functions will make testing easier and at the same time will make the tested code better (= easier to understand and so easier to maintain and reuse).
Again — a definite win-win, thanks to the use of unit tests.
I hope I managed to explain the meaning and importance of good code and that you agree that becoming a better coder it is a goal worth pursuing.
The tools that I introduced above were created for and are used by teams of software developers to enable better collaboration, better standards and better quality.
However, as I hopefully demonstrated, these tools also have massive value beyond their mainstream use as aids for self growth and improvement for individual developers and is small projects.
These tools, when used wisely allow us to interactively and incrementally learn best practices and produce better code.
I believe that any software project, even small solo ones should use them.