Consistency through conventions

Meta Info

Consistency through conventions
Author: Stijn Dejongh
Published on: 2023-08-26 00:00:00 +0000 UTC
Ammerse Values :
A: Mi: M: E: R: S: Ex:

Problem Statement

Developers struggle to maintain code consistency, negatively affecting code maintainability and reliability.


We wish to maintain a higher level of code discipline, increasing the reliability of the code.

Contextual forces

What is this? These describe factors at play that influence the outcome of the practice. They are not necessarily good or bad, but they are important to be aware of. Some contextual forces are enablers, giving the practice a higher chance of being useful. Others are deterrents, making the practice less useful in your given context.


  • Your language of choice has a recommended style guide
  • Automated code analysis/linting is in place, or available
  • Your team is cohesive and coherent, it is able to align on a preferred style
  • Your development environment offers a way to share/import code styles


  • We have a need for expressing our individuality.
  • You work with many distinct codebases or programming languages.
  • Manually reviewing code for consistency is a tedious task, that most prefer to avoid.
  • Normalization of deviance: if most of your operating process is “exceptional”, the “exception” becomes the norm.


Define a consistent coding style for your team, document it, use it, and enforce it. If possible, add automation to your build process to point out any irregularities.

Even if code is not fully up to spec, it is better to be consistent with the rest of the surrounding code than to introduce yet another style to an existing project. If you are going to change conventions midway through the life-cycle of a codebase, be sure to change all of it. Do not let inconsistent styles creep into your repositories.

Being consistent within a codebase does not mean your need to apply the same style to all repositories within your project. It can be a viable option to apply a certain style within one repository, and a different one in another. As long as the code is consistent with it's surrounding, you will get the benefit of pattern recognition.


  • Code is write-once, read-often.
  • Working with consistent representations lowers mental load.
  • Humans are excellent at recognizing deviations from the norm.


  • Developer Resistance: Enforcing a uniform style might face pushback from developers who value their coding individuality.
  • Initial Overhead: The process of defining, documenting, and implementing consistent coding standards can be time-consuming and resource-intensive.
  • Tool Dependence: Reliance on automated tools for enforcement might lead to over-dependence, reducing manual oversight and potentially missing nuanced issues.
  • Stifled Innovation: Strict adherence to coding conventions might stifle creative problem-solving and innovative coding techniques.
  • Initial Overhead: The process of defining, documenting, and implementing consistent coding standards can be time-consuming and resource-intensive.
  • Power Dynamics: Senior or vocal team members might dominate the creation and enforcement of rules, creating an oppressive atmosphere for others, particularly new or timid team members.
  • Onboarding Challenges: While having a set of conventions aims to reduce variation and cognitive load, overly strict enforcement of coding conventions might overwhelm new team members, making it difficult for them to integrate smoothly.

Mitigation Strategies

  • Inclusive Decision-Making: Involve the entire team in the process of defining the coding standards to ensure buy-in and reduce resistance.
  • Incremental Implementation: Introduce the standards gradually, starting with the most critical rules, and expand over time to allow for adaptation.
  • Flexible Enforcement: Allow for exceptions where necessary and encourage developers to suggest improvements to the conventions.
  • Comprehensive Training: Provide training sessions and resources to help developers understand and adopt the new standards effectively.
  • Regular Reviews: Conduct regular code reviews and discussions to address any issues with the conventions and make adjustments as needed.
  • Inclusive Decision-Making: Involve the entire team, including junior and timid members, in defining coding standards to ensure broad buy-in and reduce the risk of power imbalances.
  • Mentorship Programs: Pair new team members with mentors who can help them understand and adapt to coding standards without feeling intimidated.


Nuke ’em, duke!

A great example of this is the use of brackets to identify code blocks in Java. While not strictly necessary to make your code compile or work, many teams explicitly ask their developers to include brackets, even on conditionals that could be written as one-liners. This particular convention tends to irk newer developers who like their code to be as concise as possible. While brevity is a usually good, at times it can confuse your readers.

Let’s have a quick look at the code below, which checks whether we are under attack, and whether to launch a retaliatory nuclear strike.

public void failSafe(boolean thePresidentIsDead){


Now imagine someone in the team is tasked to add logging statements to the entire codebase. They do as they are told, resulting in the code below. At first glance, everything looks fine. Can you spot the cataclysmic mistake?

public void failSafe(boolean thePresidentIsDead){"Checking to see if everything is still okay...");

        LOG.warn("OMG! You killed him. You bastards!");
        launchNukes();"Carry on, carrying on");

The nukes will be launched, regardless of the president being alive or not. A simple log statement, combined with misleading indentation has set forth a thermonuclear apocalypse.

Use Case: Enforce coding standards automatically

Most (semi-mature) languages have a preferred style. It is usually wise to stick to it, even if you do not personally like it. As a few examples: Java is written mostly in camelCase, C# uses PascalCase for everything but variables, Python and JavaScript developers tend to put expected variable types in the names. The list on small and large differences between different language default writing styles is long.

Even on your project, you might decide to use a particular spacing/indentation convention, method ordering, or you might enforce a certain level of test coverage for parts of your project. Basically: there are a lot of custom conventions that exist within a project. On most projects, the biggest part of review comments are related to people not following the preferred style correctly.

In order to avoid mostly useless back-and-forth on stylistic differences, delegate these things to an automated system. There are plenty of good linters and static code analysis tools out there that let your automatically find and highlight issues with the style of code that is being written. Additionally: most modern IDE’s allow developers to import a stylesheet into their tooling, and reformat the code prior to committing it to the source control system.

tip: Let a machine handle the easily spotted and corrected “issues” in your code base, enforcing consistency. Spend your time on digging into the important things, such as functionality or the architecture of the code.