Before we start - I'm working on https://cloudash.dev, a brand new way of monitoring serverless apps 🚀. Check it our if you're tired of switching between 50 CloudWatch tabs when debugging a production incident.
Time flies, doesn't it?
My programming journey began in 2012, with my very first C++ internship. Frankly, I had absolutely no idea what I was doing (this hasn't really changed). Nevertheless, I've picked up some lessons along the way.
Disclaimer: There isn't going to be any code whatsoever in this post.
Question: What is the most important language in programming?
It's English.
Or Spanish.
Or Chinese.
Or Polish.
Or whatever you use to communicate with other people at work.
Talking to humans is way more important than talking to machines
Programming is a team sport. On rare occasions you might see a brilliant product built from scratch by a single person but in the vast majority of cases - you need a team.
Communication skills can make or break a project. Don't worry, it's not just you and your team, NASA is struggling with that as well.
Soft Professional skills can be more important to a project success than purely technical ones. Who cares if you hire 5 of the best database experts in the world if they refuse to talk to each other and you end up with 5 different instances of MySQL, Aurora and MongoDB.
Have a deep understanding of what you are building and why
Most people are happier when they have a sense of purpose. This applies to work as well.
As a software developer your goal is not to translate JIRA to JavaScript, Trello to C# etc.
Your goal is to solve problems with code.
If you have a deep understanding of the system you're building/maintaining then you can make decisions outside of pure tech. Is this feature even necessary? What problem does it solve? Can we solve this problem any other way? Do we want to solve this problem in the first place?
This line of thinking is sometimes referred to as business context, but if you want to do your job well, you should not only understand the context, but to be able to shape and influence that. You don't have to have a C-level position in your organisation to influence your product. Or at least - to understand it.
If code review in your team is a stressful experience you are doing it terribly, terribly wrong
Oh boy. Code review.
We really don't think about it but the act of putting our work out there in public and have it reviewed by multiple other people is a bit unique to our profession. No wonder people can be anxious about the whole experience.
I have personally seen people submitting code reviews when X wasn't in the office, or Y was at a business trip. X was a brilliant programmer but enduring through his code review process was a chore. If you leave 50 nitpicky (is that a word?), unkind comments under a PR of someone who is a junior programmer, you are not proving your superiority as a developer. You are proving that you're not a good human being.
Okay, but what do I do when I see that this feature is completely broken?
Stand up. Reach out to that person in private. Talk to them, find out why they implemented that code this way.
Most people do not want to write bad code. And if they do, they probably are dealing with constraints you're not aware of. They could also not be really good at programming (yet) and it's your opportunity to shine as a mentor.
Something WILL go wrong, be prepared
According to wikipedia:
Murphy's law is an adage or epigram that is typically stated as: "Anything that can go wrong will go wrong".
It's one of those things that are too true. Always assume that something may break when designing a system.
If you're building a login form, assume that people will copy&paste an entire book into the password field.
If you're building a WYSIWYG window, assume that someone will try to break it, and they are likely to succeed.
If you have a database, it will go down at some point. If you haven't tested recovering your database from a backup, it's not a backup.
If you're doing a live demo in front of an audience - make sure that the demo works online, offline, upside down and under water.
Don’t be afraid to say “I don’t know”
The best part of having a senior next to my job title is that I can finally respond to a question saying:
I don't know, never tried that. I'll take a look and I'll get back to you.
When I was a junior, I was terrified of someone figuring out that I'm a fraud. After a couple of years as a developer - if I haven't seen something, it could be that it wasn't relevant till now. Or I just have another cool piece of tech to learn. Lifelong learning is not a buzzphrase in software development, it's the reality.
Or I'm just an incredible fraudster, managing to fool all those people that I can actually do my job. You never know.
Learn in public
Once you go from "I don't know" to "Okay, that was interesting" - share that with someone. Write a blogpost, record a video, do a talk at a company knowledge sharing event or just ... tell someone. If you think that something is obvious to everyone, it's not. Even the most senior people have something to learn from beginners and vice versa.
Teaching is an incredible way of ensuring that you really understand the subject in question.
As the saying goes:
When one teaches, two learn - someone hella smart
What are your lessons learned as a developer?
Oldest comments (45)
This was exceptional. Thank you for sharing.
"Programming is a team sport." I've been saying this exact thing for years. Thanks for sharing your experience it was a really good read 😁
In a typical corporate environment? For sure, but I certainly know quite a few of programmers who are very successful developing solo.
"Stand up. Reach out to that person in private."
"Learn in public"
So which is it? Okay, if we take the context into account it's not as contradictory as here, but still it could have been emphasized more in the article that the first applies to a specific situation, whereas the second is a general rule.
This was amazing. Loved it!!
This was a really great article! Thank you for sharing your learnings Tomasz!
I really liked that you focussed not on code or technical factors, but rather on the human aspect of working as a software engineer.
Having said that, I'd love to read another part of this story where you write about your learnings from a technical point of view. Things like how has your attitude towards problem solving, coding choices, etc. changed over time. Or any other learnings, in fact :)
I'm sure the community would benefit a lot from reading about your journey and learnings.
Pure gold, thanks for writing this!
This is fantastic. Thank you so much for sharing.
Excellent read! I couldn't agree more with these, I have learnt the same lessons in the past 5 years, often the hard way. Murphy's Law, as you said, often finds a way of teaching you to be more thoughtful in future.
I loved the article and agree with almost all of it, particularly the "things go wrong". I feel that you need to switch over periodically and become your own enemy and try to think of obnoxious things that could conceivably happen. I remember we had a QA guy named Dan Song - an ace. Once he bet us he could immediately break this program we were working on - one he hadn't even seen - and we said, "Sure". He walks over, we bring up the program, he hammers on the keyboard incredibly fast, it crashed instantly. The other programmers were all, "That's not fair!" but I had a revelation.
And Dan had a lot of tricks other than that. I realized that before then I had been testing code to try to show it working - now I started testing my code to try to make it break. Changed my life!
The one thing I disagree is the "50 comments on a junior developer's first review makes you a bad person". I disagree - it's doing this out of the blue, with hostility that makes you bad.
I started in a new organization with 20 years' experience under my belt and the first five code reviews were relentless with huge numbers of comments! But all the comments were warm and friendly, and I had been warned in advance that it took a long time to get all the details right in this huge system. People were supportive and told me that I was doing a good job. I was told to expect perhaps hundreds of comments on my first significant code review, so when there were only about 50, I felt I was doing OK.
Reviewing code in private is wrong - because you want to share the information with the team. I learn more from reading other people's code reviews than perhaps anything else. I ask myself, "Why did they do this weird thing?" and then I read the code review and I say, "Because it's the only way that will work!"
Any number of comments on a code review are fine - if the reviewee is primed to expect this before anything happens, and if the comments are supportive and positive.
Great article though - this is just a quibble.
There is a reason why SO is so beneficial to the Developers. Many people ask questions(newbie, stupid, or smart) on that platform and the good peer reviewed comments from the experts is documented right there. That encourages people to learn in public and share their knowledge.
Similarly, I prefer the discussions to happen in a common place such as a GitHub PR or on the related GH/Jira issue. Team should be encouraged to take a look at them and have an understanding of the reasons behind a specific implementation.
The problem might be the mindset of the team members who think that asking too many questions about an implementation means it's bad code or the implementer isn't skilled enough.