AI, Ethics, Programming languages, Health, Diversity and Salaries
Yesterday StackOverflow released this year’s data from their annual Developer Survey.
Their recap is always interesting and full of trends and helpful stats, but it’s also quite long. Not all of us have the time to sift through 60+ pages of bar charts and percentages.
Fortunately, I do.
I know I’m a dev and should go digital, but I still do love to just use a pen & some paper
Having carefully read over it a few times, here are the main things to take away from this year’s stats.
Let’s start with a confirmation of an intuitively known fact:
Code — Developers love it.
This one makes a lot of sense but a reminder never hurts. Developers are not (only) in it to make a living, they seem to truly enjoy programming as over 80% of developers code as a hobby regardless of their job status.
Actually, the developers with the less time to devote to it (because they exercise daily, spend more time outside or have care-taking responsibilities) seem to be those who code the most.
To add to that**, every other developer also contributes to open-source software** (OSS), although some languages seem to be especially favored by OSS contributors, such as Rust or Clojure, with 70% of its developers being OSS contributors.
To confirm this trend even more, over 3/4th of all developers say they go to Hackathons is simply that they enjoy them, and half say they go to them to improve their programming skills. Which is coherent with the fact that 87% developers have learned a technology or programming language by themselves.
Point is, we love to code, and we code a lot.
Health — Bad habits, good furniture.
Our love for code might have an impact on our health though, as 1 out of 3 developers will skip at least a meal per week to be more productive.
But despite our bad habits, 1 out of 2 developers uses a standing desk and just as much have an ergonomic chair, ergonomic mouse.
Tech — What’s in fashion?
That’s where the money is after all (although you might say that’s more applicable to the upcoming Salaries section). So let’s get to it.
Honorable mentions
- Rust sails through its 3rd year in a row as Most Loved Programming Language
- Visual Basic 6 (VB6) through its 3rd year in Most Dreaded Programming Language
JavaScript is here to stay — and we seem happy about that.
JavaScript is the most used Programming Language once again, it’s also the 2nd language that developers want to try the most.
NodeJS, React and Angular (the 3 most used frameworks across all languages) place 1st, 2nd and 4th respectively as frameworks that developers want to try the most and people seem to be right to want so, since they place 2nd, 4th and 7th on the list of the most loved frameworks.
MongoDB, a regular part of most JS stacks, is the database most developers want to try.
Redis & PostgreSQL are the darlings of the database world.
MySQL and SQL Server are the most used databases, with PostgreSQL coming up as third. But PostgreSQL comes up as 2nd most loved database technology, and 3rd most wanted to try.
It’s in good company**,** as Redis takes 1st place for most loved for the second year in a row and 4th as most wanted to try.
Python still rises
Python appears to be the fastest-growing major programming language. After dethroning PHP of its #8 spot last year, Python now takes over C# for the spot of 7th most used language.
For the second year in a row, Python is the language most developers want to try. It’s also the 3rd most loved.
Its Machine-Learning framework, PyTorch, is the 3rd most loved framework.
TensorFlow dominates the Machine Learning world
TensorFlow is the 3rd framework most developers want to try. And not without reasons since it’s the top, most loved framework for the respondents.
AI — Overall, we’re excited.
Concerning AI, it appears that as developers we’re more excited than afraid.
The survey studied four aspects of AI:
- The automation of jobs
- AIs becoming smarter than humans (the singularity)
- The definition of “fairness” evolving in human/algorithmic decisions
- Algorithms making important decisions
And as it turns out we don’t seem to find one prospect to be scarier than any other. But as could have been expected from the masters of automation we’re mainly excited about the automation of jobs over the others (40% of developers find it exciting).
We also mostly tend to think that studying the implications of AI and its ramifications is a task that should befall on the people designing and building it. Only 1 developer out of 4 thinks that it should be a governmental/regulatory body’s responsibility.
Ethics — We have a moral code, but it still depends.
95% of developers would hesitate before writing clearly unethical code. 58% would flat-out refuse.
95% of developers would report it depending on the situation. 13% would do so publicly.
This is a good news for our moral integrity, with only 1 developer out of 20 being ready to go against what they think ethical, but within the remaining 95%, about 35% would adjust their behavior to the situation, no hard and fast rules in ethics for developers but 80% of us think that we have an obligation to consider the ethical implications of our code.
Ultimately though, most of us (57%) think that management is responsible for them.
Diversity — Coming up!
The next generation of developers still won’t be quite the melting pot, but we’re getting closer.
1 out of 3 programming students are people of color, where only 1 out of 4 professional developers are.
The increase is especially noticeable for students of Middle-Eastern (148% growth) and East Asian (145%) descent.
Sexual orientation and gender also see some diversification when looking at students and developers with little experience:
Where 6.6% of developers identify as Gay, Lesbian, Bi or Queer in the professional world, 8.3% of student developers do so.
The rate of women in the industry also rises from 6.6% to 7.4% between professional developers and students developers.
The number of trans and non-binary individuals in the community is also on the rise when looking at the student population. Interestingly, these also tend to contribute 1.5 to 2 times as much to open source than their cisgender peers.
Building a diverse company
Where men seem to value compensation and benefits as the main criterion to pick a job, minorities seem to value the company culture first.
More specifically:
- Women value the company culture, the opportunities for personal development and the technologies used about equally (16.9%, 16.8% and 16.4%)
- Trans and non-binary people value the company culture way more than any other criterion (5–6% lead over the second most important)
- Cisgender developers consider the diversity of the company to be the lowest priority (out of seven possibilities) when assessing a potential job, whereas it’s the 4th most important for both non-binary and trans people.
Salaries — Follow the money
We’ve all seen float many articles about how “X is the language to learn in 2018 if you want to make the big bucks.”. Let’s see how those hold up next to the data:
The Top 5 programming languages with the highest-median salaries this year are: (in USD)
- F# — 74K
- OCaml — 73K
- Clojure — 72K
- Groovy — 72K
- Perl — 69K
- Rust — 69K
The Top 5 positions with the highest-median salaries are: (in USD)
- Engineering Manager — 89K
- DevOps — 72K
- Data Scientist — 60K
- Business / Data Analyst — 59K
- Full-Stack Developer — 59K
The position of Data Scientist is notable, as it is one of the highest median salaries yet has one of the lowest average number of years of experience. (60K for an average of a bit under 6 years of experience).DevOps are more experienced but are still more compensated on average than their non-DevOps peers (72K for about 8 years of experience on average).
These are the statistics and takeaways I’ve found the most interesting from this year’s report. I have no doubt some other fact and stat I’ve decided not to include here might be of importance to you and I strongly advise reading through the original report when you find the time to do so.
I hope my digest was enough to give you a comprehensive view of the current state of the industry as of 2018.
If you’re one for data points I’d advise you to wait for a few weeks until StackOverflow publishes the whole dataset, I have no doubt more detailed articles will flourish then.