“Difficult to make even small changes”
“Breaking functionality of software by making changes”
“Introducing a new bug by fixing another one”
“Implementing code that is unnecessary”
“Almost impossible to add a new feature because of complicated code”
“The never-shipping product”
“Throwing code away and rewriting it from scratch”
Are all the above statements familiar?
Every minute a developer from any part of the World says (or thinks) any of the above statements and wishes to cry. Why?
These are common issues talked about by developers very often. These stories are experienced in every development team.
There are many small factors that are slowly and gradually harming the developers’ projects. They are not immediately destructive. Most of them only do long-term damage. Something you won’t see the damage for a year or more. So when somebody proposes them, often they sound harmless. Even when you start implementing them, they may seem fine. But as time goes on — and particularly as more and more of these stacks up — the complexity becomes more apparent and grows until you’re another victim of that ever-so-common horror story.
To avoid being one of the victims, you should embrace the fundamental laws of software. You should develop a mindset that every developer should have. This mindset will help you make better decisions in your daily programming journey. You can keep your software as simple as possible. You can protect it from being an unmanageable and complex system.
Here are the key points that every developer must master.
1. Conceiving The Purpose of Software
First of all, you should understand the purpose of the software. There is, in fact, a single purpose of all software: To help people.
Remember: the purpose of software is not to show off how intelligent you are.
Developers who cannot conceive the purpose of software will write bad software. What is bad software? A complex system that doesn’t help people that much.
When you are making decisions about software, you should guide yourself by always keeping this in mind: How we can help? You can even prioritize feature requests this way.
2. The Goals of Software Design
Every programmer is a designer.
When software is hard to create or modify, developers spend most of their time focusing on making things “just work,” and less time focusing on helping users. The design of software aims to make developers job as easy as possible so they can focus on what matters. You will create software that will help users and your software will continue to help them for a long time.
However, if you design a bad system, your software’s lifetime will be short.
This brings us to the most important goal of the software design:
To design systems that can be created and maintained as easily as possible by their developers, so that they can be — and continue to be — as helpful as possible.
So there are two key points here: Your design should be easy for you and helpful for others.
Developers who don’t fully understand their work tend to develop complex systems. It can become a vicious cycle: misunderstanding leads to complexity, which leads to further misunderstanding, and so on.
Actually, one of the best ways to improve your design skills is to be sure that you fully understand the systems and tools you are working with.
Understanding is the key difference between a bad developer and a good developer.
Bad developers don’t understand what they are doing, and good developers do. It really is that simple.
Simplicity is the ultimate sophistication. — Leonardo da Vinci
Programming is the act of reducing complexity to simplicity. A “bad developer” is just somebody who fails to reduce complexity. A “good developer” is doing everything in their power to make the code as simple as possible for other programmers.
A good developer creates things that are easy to understand so that it’s really easy to shake out all the bugs.
Now, developers are generally intelligent people and none of them likes to be treated like they are idiots. Ironically, this leads them sometimes to create things that are a bit complicated. They basically think like this:
“Oh, other developers will understand everything I’ve done here. I should write some clever code that is hard to understand so that they can think that I am very smart.”
A mistake caused by a wrong mindset-not necessarily by a lack of programming skills. Most of the programming failures happen because of that mentality.
Showing off that you are smart doesn’t help them.
Developers who are new to your code don’t know anything about it; they have to learn.
So, you should ask this question: “Do I want people to understand this and be happy, or do I want them to be confused and frustrated?”
The truth is that if other developers who read your code can understand it easily it means that you are doing good.
Complexity has nothing to do with intelligence, simplicity does. — Larry Bossidy
The question is: “How simple do you have to be?”
Here is your answer: Stupid, dumb simple.
Controlling complexity is the essence of computer programming. — Brian Kernighan
The source of many software failures is complexity. You start out with a simple project that can be completed in one month. Then you add complexity, and the task will take up to three months. Then you start to add features that fulfill some other purpose. Things get very complex because you expand your software purpose for no reason. The tasks will take six months.
But that is not the end.
Then you take each piece of the feature and make it even more complex, and the task will take nine months. Then you start to introduce many new bugs because of the complexity in your code. Naturally, you start fixing them all without thinking how these fixes will affect other parts. At the end, when even small changes become hard. When bug fixes start to introduce new bugs, you will come to one of the most popular programming horror stories: Rewriting code from scratch.
So, how did you become a victim of this horror story? Nah, who cares. It’s better to ask: How could you avoid being a victim?
Well, it is simple. First, you will exactly know your software purpose and its definition. Second, you will be as simple as possible in every piece of code you write. Third, when a new feature or change request comes to the discussion table, you will evaluate them based on your software purpose and question them.
As a developer, your first behavior should be resistance to (unnecessary) change. This will prevent you from adding unnecessary codes into your software. When you are convinced that this change is a need, then you can implement it.
There are many factors that will increase complexity but those are the most popular ones. Aside from everything, there is only one rule that you should follow:
Your main purpose is to control complexity, not to create it.
Maintenance is one of the most important things in software development. Unfortunately, developers usually ignore how important it is. Quick coding and fast shipping look more important than code maintenance. This is the point where they make a mistake — ignorance of future code maintenance.
There will always be some implementation of changes. Not only you have to implement them, but you also have to maintain them over time. As a developer, thinking about future maintenance of changes is one of your main responsibilities.
All changes require maintenance.
Simplicity and complexity are the two main factors that affect code maintenance. The ease of maintenance of any piece of software is proportional to the simplicity of its individual pieces. The effort of maintenance is proportional to the complexity of the software.
The one rule that you should follow about maintenance is:
It is more important to reduce the effort of maintenance than it is to reduce the effort of implementation.
Consistency is a big part of simplicity. If you do something one way in one place, do it that way in every place. For example, if you name a variable thisIsVariable, then all of your variables should be named that way (otherVariable, anAnotherVariable, etc. not other_variable).
Code that isn’t consistent becomes harder to understand. Don’t keep forcing developers to relearn the way your system works every time they look at a new piece of it.
In any team sport, the best teams have consistency and chemistry. — Roger Staubach
How do you make decisions about your software?
When you face many possible directions, how do you decide which option is the best? What to focus on and which features you should implement?
To answer those questions, there are three important factors that will help you make a better decision:
- The desirability of a change (D): How much do you want to do the change?
- The value of a change (V): How much value does this change bring or how much does this help your users?
- The effort required to perform the change (E): How much work will the change require?
The equation is simple: D=V/E
The desirability of any change is directly proportional to the value of the change and inversely proportional to the effort involved in making the change.
When you prioritize your work, you should follow this rule:
The changes that will bring you a lot of value and require little effort are better than those that will bring little value and require a lot of effort.
9. Solving Problems
The first step is understanding. Know exactly what is being asked. Most hard problems are hard because you don’t understand them. Write down your problem and try to explain it to someone else.
If you can’t explain something in simple terms, you don’t understand it. — Richard Feynman
The second step is planning. Don’t take action. Sleep on it. Give your brain some time to analyze the problem and process the information but don’t spend too much time on planning.
Think before acting.
The third step is dividing. Don’t try to solve one big problem. When you look at the problem as a whole, it can scare you. Divide it into smaller tasks and solve each sub-problem one by one. Once you solve each sub-problem, you connect the dots.
10. Good enough is fine
“Perfect is the enemy of good.” — Voltaire
Whether creating a new project or adding a feature to existing system developers tend to plan everything out in detail from the beginning. They want the first version to be perfect. They don’t focus on the problem they will solve and how their software will help people. They start by thinking every small detail they could think. Then assumptions and predictions come along followed by “What if” sentences. They have to predict the future because they were now so captivated by the imagination of the project in their mind and their project has to be as perfect as they imagined it.
Actually, they are not aware of what’s waiting for them and how much it will cost them by chasing perfection.
Let me tell you what will happen:
- You will be writing code that isn’t needed
- You will increase complexity by adding unnecessary codes
- You will be too generic
- You will be missing deadlines
- You will be dealing with many bugs caused by the complexity
Do you want this to happen? I guess no.
What you should instead?
Start small, improve it, then extend.
Incremental design should be your guide. Here is how you would use it to design a calculator:
- Plan a system that does only addition and nothing else.
- Implement it.
- Improve the now-existing system’s design so you can add other operations also.
- Plan subtraction and repeat step 2 and 3.
- Plan multiplication and repeat step 2 and 3.
- Plan division and repeat step 2 and 3.
“A prediction is simply a forecast that something will happen in the future. It could be factual and based on some kind of objective data or it could be based on an assumption.”
When faced with the fact that their code will change in the future, some developers attempt to solve the problem by designing a solution so generic that (they believe) it will accommodate to every possible future situation.
Being too generic involves a lot of code that isn’t needed.
You can’t predict the future, so no matter how generic your solution is, it will not be generic enough to satisfy the actual future requirements you will have. Most probably, this time will never come and the code you wrote to solve future problems will increase complexity, make it hard to change the pieces of code and eventually it will become a burden that may destroy your software.
Don’t predict to future. Be only as generic as you know you need to be right now.
What is the assumption?
“An assumption is something that you accept as true or suppose to be true, although you have no conclusive proof.”
One of the great killers of a software project is assumptions. Let’s see how an assumption can kill a software project.
A developer knows that they have to develop a system to do X. Then they think that the system will require them to do Y in the future, and they implement Y as well. They write thousands of lines of code to design Y.
In the future, the developer realizes that the current requirements are completely different than what they thought. But now, the software has unnecessary codes that make it hard to throw away because everything is intertwined. It takes months to refactor the code and now they think to rewrite the whole software from scratch which will cause them to lose months.
To avoid being a victim like this developer, follow this simple rule:
Code should be designed based on what you know now, not on what you think will happen in the future.
13. Stop Reinventing
If, for example, you invent your own garbage collector when a perfectly good one exists, you’re going to be spending a lot of time working on the garbage collector, when you could just be working on your software.
The only times it’s okay to reinvent the wheel is when any of the following are true:
- You need something that doesn’t exist yet
- All of the existing “wheels” are bad technologies or incapable of handling your needs
- The existing “wheels” aren’t being properly maintained
Don’t reinvent the wheel.
As a developer, your first reaction to changing requests should be “NO’’.
Always resist adding more code, more features until you are convinced that they are required and there is a need to implement them. Because unnecessary changes will increase defects in your software.
How can you know that there is a need for them?
Go back and remember your software purpose. Then remember the simple equation in prioritizing section.
From: email@example.com (Russ Cox)Subject: Re: [9fans] design clairvoyance & the 9 way
Date: Thu, 8 May 2003 04:05:31 GMT
> What does tomorrow's unix look like?
I'm confident that tomorrow's Unix will look like today's Unix, only cruftier.
Don’t spend your time on repetitive tasks. Set them up and forget about them. They can work while you are sleeping. When you realize that you are doing something again and again, just remember this rule:
If you can automate it, automate it.
16. Code measurement
Measuring programming progress by lines of code is like measuring aircraft building progress by weight.
— Bill Gates
I see developers who measure their software quality based on code lines. They think that more code lines mean that they are doing a great job. The software contains hundreds of thousands of lines of code, which means the software they work on is so big.
The question that pops up here is: Is it really that big, or there is something wrong there?
The answer is that most probably there is something wrong with their design. Most of the simple solutions don’t require a lot of code. You can achieve simplicity with a little bunch of code and solve the problem.
I’m not saying that the least code is always better. While you want to avoid having less code, you can easily fall in a trap that will cause you to write clever code that is hard to understand for others. You should find a balance.
The optimum code is a small bunch of code that is easy to understand, easy to read.
How do you measure your productivity?
By writing more lines of code or by throwing hundreds of lines of code away?!
Your main goal should be keeping your code base as small as possible. The question is not “How can I write more code?” rather it should be “How can I remove more code?”
“One of my most productive days was throwing away 1000 lines of code.” — Ken Thompson
When should you add logging and error handling to your project?
You should add logging in a very early stage. This will help you to find the problem easily and save your time.
I see many mistakes when it comes to testing code. Let me give you an example. There were two conditions, a simple if-else block. The developer gave input to the software which will enter inside the if block. They tested it and committed code to source control. Done! What about else block? When software was shipped to production, that caused a lot of errors. When you test your code, you must execute all new lines at least once and you should start to test parts before the whole.
When you have a bug, first you should reproduce it. You shouldn’t guess the source of the bug and apply fixes based on your assumption. Most probably, you will be wrong. You should see it with your own eyes before applying the fix.
You should be reliable. When other developers in your team see that you committed new code to source control, everyone should know that your code is tested, and works.
Untested code is the code that doesn’t work.
Developers’ estimation sucks.
Usually, they underestimate things rather than overestimate them. They underestimate the time and effort required to develop a small amount of code or a feature. In the end, this underestimation leads to missing deadlines.
The solution: Break the big thing into smaller things. The smaller it is, the easier it is to estimate. You’re probably still going to get it wrong, but you’ll be a lot less wrong than if you estimated a big project.
Everything takes longer than you think.
20. Running Away From Rewriting
I believe that when you embrace the fundamental principles of software development mentioned in that article, you won’t come to this point. However, if, somehow you make these mistakes and find yourself thinking about rewriting your code, here is the main thing that you should know:
Rewriting code is often a developer delusion, not the solution in most cases.
Why is it a delusion?
Well, because it’s harder to read code than to write it. This is why it is so hard to reuse code. This is why our subconscious mind whispers to us “Throw it away and start over” when we read another developer’s code.
There are many cases that you should consider to rewrite your code from scratch and you can read them here. But, here is simple advice for you:
Refactoring should be the first option.
21. Documentation and Commenting
One of the common misconceptions about commenting is that developers add comments that say what code is doing. This is wrong. That should be obvious from reading the code. If it’s not obvious, it means that it is not readable and it should be made simpler.
When you can’t make the code simpler then you should add the comment to explain this complexity.
The real purpose of comments is to explain “WHY” you did something, not “WHAT” the code is doing. If you don’t explain this, other programmers may be confused and when they go to change your code they might remove important parts of it.
Write a comment to explain “WHY”, not to explain “WHAT”.
Another thing is documenting. It is important to have documentation to explain your software’s architecture and every module and components. This is required to see the high-level picture of your software. When a new developer joined your team, it would be easier to understand the software as a whole. When developers don’t have any clue about other parts of the software, they could easily make a mistake in their own part which can affect other parts also.
22. Picking Technologies (Tools, Libraries, etc.)
First things first, always remember this rule:
Don’t depend on external technologies or reduce your dependency on them as much as possible.
Why is that? Because they are another common source of complexity. They can kill your active development and make everything even harder. When you are dependent so much on external technologies, you are not free. What if there is a major bug in that technology? You have to wait for the developers to fix that bug and if this technology is in the center of your project basically you are stuck, you can’t move forward. So that’s why it is so important to pick the right technologies for your project.
There are a few factors you should consider before you start using some technology:
- Is there active development behind it?
- Will it continue to be maintained?
- How easy is it to switch away from?
- What does the community say about it?
If you can find the right answer these questions, you can reduce the risk of picking the wrong technology.
Keep learning. Try out different programming languages and tools, read books on software development. They will give you another perspective. Every day small improvements will make a real difference in your knowledge and skills.
Be open-minded. Don’t be obsessive about one technology. Use the required technology to solve a specific problem. Don’t be in the unnecessary discussion like Microsoft vs Linux:)
Know that every specific problem has its own specific solution.
24. Don’t be a hero
A lot of times it’s better to be a quitter than a hero.
For example, let’s say you think a task can be done in two hours. But four hours into it, you’re still only a quarter of the way done. The natural instinct is to think, “But I can’t give up now, I’ve already spent four hours on this!” So you go into hero mode. You’re determined to make it work (and slightly embarrassed that it isn’t already working). You grab your cape and shut yourself off from the world.
Don’t be obsessive. Know when to quit. Don’t hesitate to ask for help.
25. Don’t Ask Questions… Ask For Help
When you have something to implement and you are not sure about the solutions, don’t ask others how to so it …at least not immediately. Instead, try anything and everything you can think of. This is more important the less comfortable you are with a concept or language.
When you can’t think of anything on your own, search! Find answers and try them out. Modify those answers, see if you can understand why they work, adapt them to your code.
…But always seek advice.
When you have tried everything, and preferably after you have a working solution, now is the best time to seek advice. Look to peers and senior developers to review your code.
Thanks for reading! I hope this guide helped you out!
Originally published at blog.huseyinpolatyuruk.com.