Do things REALLY change in Software?

Or: does everything move at the same pace in SW? (Image property of CA Technologies)

If there is one domain in business and technology that has a reputation for dynamism and fast-paced attitude, that’s Software.

For (at least) the last two decades Software has shaped the world we live in, starting from impactful technologies (does anybody remember a world without smartphones?) to business relevance (just take a look at the list for the Forbes world’s 100 richest). New paradigms for how to run a successful business have appeared, with concepts like Agile becoming mainstream and promoting a company culture focused on innovation and change.

But does this apply everywhere? Is it possible that even the Software world has its own “pockets of resistance” to change, and in case, what would they be?

Let’s take a look at some data

To get some quantifiable insights, I decided to leverage the StackOverflow survey data.

StackOverflow is a (very) popular site providing a community for developers all over the world to ask and answer (mostly) technical questions: it’s been around for a while and in the last decade or so has presented its users with a survey to better understand them. The population of respondents has grown noticeably in the last few years, reaching almost 100000 people in 2018:

StackOverflow surveys — Number of respondents vs. Time

The questions in these surveys have changed through the years, but some things have maintained consistency, and they could provide some indication of where things are moving faster and where… not so much.

Note: All the code I used to look at the data from the surveys can be found here.

Technology first!

Coming from a forum targeting developers, it’s not surprising that a significant portion of the questions has always been devoted to technology: what kind of languages/framework/Operating System has the respondent used in the previous year, what would most likely be used the next one and so on.

If we take a look at the data for the last 5 years we see some patterns, for example in the languages used:

StackOverflow surveys’ results — Languages used (%)

It’s clear that JavaScript and SQL are a consistent presence in the lives of the respondents, and we can also see how Python seems to gain some traction.

This could also be seen from the data about the desired languages for the same years, where we find Python consistently up top:

StackOverflow surveys’ results — Languages wanted (%)

What else?

Besides technology, the survey has also traditionally tried to ask demographic questions, like the Country where the respondents live, the gender they identify with and, most recently, their ethnicity.

The question about the gender, for example, has changed through the years giving the respondent more options to better self-identify. However, if, for the sake of simplicity, we focus for a moment on the traditional binary answer (Male/Female), that in any case accounts for the vast majority of the data, something comes up pretty evidently:

StackOverflow surveys’ results —Gender of the respondent (%)

The population of respondents to the StackOverflow survey is, for more than 90% composed by men. Even more significantly, and more in line the initial question, the numbers do not seem to change too quickly: I added some simple (linear) trend curves, that seem to approximate fairly the actual data, and if we just extrapolate this model into the future we can see that a more even distribution of genders is not exactly behind the corner:

StackOverflow surveys’ results — Gender of the respondent (%), extended trend line

Some thoughts

Of course, the content here does not have any pretense of being comprehensive or conclusive. However, the numbers are pretty striking and, based on them, it seems unlikely that, at the current pace, achieving more representativeness for today’s minorities would happen quickly.

Personally, I found this confirmed at least by with my own experiences: I’ve worked as a developer and as an architect in many teams, companies, countries, and I’ve pretty much always experienced a skewed composition of people, even in places that try to be more forward-looking, like California.

I think the question for more equality in any workplace is a big, burning one for all of us today, and what I really (and simply) wanted to say here is that the system by itself does not seem to regulate quick enough, even for fast-pacing domains as Software.

So, what to do? I think that initiatives that try to break the cultural barriers that play a role in this, and promote STEM education for young women, like Girls who Code are an excellent place to start, and I truly encourage anybody who’s reading this to look for anything similar around you.

In the end, equality is not a masculine value or a feminine value: it’s a human value. Hence, yours too.

I’m a SW Architect and an Engineer, interested in IoT, Edge and Autonomous Systems