How do you manage cognitive load in programming?

Write automated tests. Automated tests, specially unit tests, provide several side benefits. By making your code testable, quality is most likely improved. Boundaries, behaviors and implementation details get refined in the pursuit of writing the tests. Also, tests serve as documentation of what your code is supposed to do. Reading tests is in many cases way easier than understanding the source code because you don't get lost in implementation details when behavior is what you care about.

Prefer using pure functions whenever possible. A pure function is one which behavior is fully defined by its arguments and which does not have any side effects (printing, changing state, etc.), and its result is fully contained within return value. Such functions are very easy to test, and they are trivially parallelizable. While using a functional programming language (like Haskell or Scala) helps, you can benefit from pure functions in most modern languages, like C or Python. Using pure functions may add some apparent overhead to your code when writing, but that pays for itself many times over when reading: you literally don't need any external context to understand their behavior!

#1: Design patterns are one thing, but following coding principles like SOLID will make you apply the more fundamental design principles automatically. #2: Convert your example scripts into unittests or systemtests, you will then have an example which shows intended use as well as a test which helps ensure intended use in one. #3: Follow a paradigm if it makes sense to do so, prefer simplicity and minimize Lines of Code over complexity and flexibility, unless adding more Lines of Code makes the code easier to understand. #4: Don't leave placeholder code in an attempt to predict the future, add the code when it's time to do so, otherwise you risk littering the code base with redundant and dead code.

Linters can help with these code metrics: - Maintainability Index. Calculates an index value between 0 and 100 that represents the relative ease of maintaining the code. - Cyclomatic Complexity. Structural complexity by calculating the number of different code paths in the flow of the program. - Depth of Inheritance. Indicates the number of different classes that inherit from one another, all the way back to the base class. - Class Coupling. Measures the coupling to unique classes through parameters, local variables, return types, method calls, generic or template instantiations, base classes, interface implementations, fields defined on external types, and attribute decoration. - Lines of Source code. - Lines of Executable code.

Before start writing your code: 1. Plan it. Take a piece of paper and pen, write down how you are going to solve a particular problem. Writing ideas on paper clarifies the clutter in mind and it also helps in reducing the feeling of being overwhelmed. 2. Research it. How world solves similar kind of problem. 3. Break tasks into chunks. 4. Work with focus. Eliminate distractions. 5. Maintain a checklist to give a sense of accomplishment when a task is done.

You can use the Single Responsibility Principle (SRP) to guide your thought process when deciding whether you need an abstraction. Another way I like to think about it is if I'm mixing unrelated concepts and, for example, working with a database, files, and domain entities in a single flow. It would quickly become unmaintainable if I used low-level code for all these. This is where a good abstraction can help reduce the cognitive load.

Abstraction can be your enemy if not applied thoughtfully and at the right time. The difference between good and great engineers is the ability to understand where and when to create the abstraction.

Managing cognitive load in programming can be challenging, but I've found some helpful strategies. First, breaking tasks into smaller, more manageable pieces helps me focus on one thing at a time. I also use comments and documentation to keep track of my thoughts and decisions, reducing the mental burden of remembering everything. When learning new concepts or tools, I take it step by step and practice to reinforce my understanding gradually. Collaborating with teammates allows us to share the cognitive load and learn from each other. And finally, taking short breaks, staying organized, and maintaining a healthy work-life balance help me recharge and keep cognitive overload in check.

Cognitive load ? You can use A.I. tools for this. No need to switch load to younger members of your team. A.I. tools are better "crunchers" and "makers", just waiting for your instructions. Effect is exponential.

Not everyone thinks in the same way. Our brains are processing information in different patterns, this is why some engineers find a specific paradigm confusing while others find it intuitive. It's helpful in a team to find a paradigm that everyone finds intuitive, therefore does not increase the cognitive load and agree in using that. For example using imperative programming, or functional or OOP.

For example, I never get used to Reactive Paradigm. I see people who are switching from React to Flutter or Angular and used that Reactive (Redux) library. Oh this Reactive people :-).

Best way to learn is to experience what other experts are doing. Follow people on github, contribute in open source projects, create your network in online communities, discuss your work with them and seek valuable feedback.

Taking a break from programming, connecting with friends, hitting the gym, and clearing your mind outdoors can do wonders. It’s amazing how often solutions come to mind when you’re not actively staring at the screen. Plus, returning to your work with a refreshed perspective and renewed energy at night can lead to breakthroughs.

Practice makes perfect *Train yourself to sene and pick up skills from outliers to critically frame your learning process. *Put coding and running the codes to see if it really works into two different time boxes.

In my personal experience it is easy to manage for a new project, where projects are divided into modules, and changing one module without affecting other modules is easier. That is mostly achieved by Microservices pattern, Devops (proper build, test, and deployment pipeline), and Automated ways to manage code format, code style, and code complexity (size of code file, number of lines in a method). The issue comes with legacy and monolith projects, where one class file contains more than 5K lines and methods with 1K lines. In that case, I go with smaller changes. Understand if the change is going to make a functional difference or not. For risky changes use a test system, deploy the change, and ensure that functionality is not broken.

Focused work (Deep Work). Sometimes this is overlooked but the most valuable work often comes from deep work. This means allocating time where you completely cut off any distractions and only focus on the task at hand. Note that some tasks allow for multitasking but most do not, especially when solving complex problems. Example of a task you can multi-task : Check/reply to your email while your code is getting compiled and deployed. Example of a task that needs deep focus : Find a potential race condition in a code segment.

Non-programming activities could contribute to a programmer's cognitive load 👇️ I've seen many corporate programmers having too many meetings to sit in. A seemingly harmless request to attend a meeting could fragment an engineer's task. It leads to more context-switching daily, causing more brain fatigue and less productive time. If you are a manager, pay attention to the daily schedule of the engineers and offer a day of the week without a meeting. As a programmer, we should all be mindful of our calendars and proactively structure the day instead of letting others fragment our day.