Making decisions - Using Statistics
This past week I attended a technology conference. One session was on Programming trends which included discussions on languages as well as coding areas. Several surveys were identified in the presentation.
The survey itself nor the size of the sample are important to this discussion. It was a very good survey. What I want to discuss is how results can turn sharply on a single metric.
So developers from all over the world contributed. 51% considered themselves to be back-end and 20% mobile developers. The analysis of the data showed that about 50% of professional developers contributed to open source but for developers using Rust, Julia and Clojure that number jumped to 70% while C#, VBA and VBNet dropped to 20%.
When I read the statement about professional developers it made me pause. Wait a minute, are these results split out between professional developers and others? The answer was yes, so that made me go back and review all of the statistics again based on this one metric.
You see on Friday night in a Uber ride from the airport to my hotel, my Uber driver told me he was developing and designing web sites and programming in php. A few days before that someone working at Sam's Club was developing solutions on the side at night while attending school. Seems like I am meeting a lot of people every day who code or consider themselves to be developers.
At this point I read the following: "Many developers work on code outside of work. Over 80% of our respondents say that they code as a hobby." What this told me is that only 20% of the respondents were professional developers. This one metric changed nearly everything in the article I had read. For example Shift and Go are growing language while Python is closing in on Java and is the fastest growing language but JavaScript continues to be the leader.
Does it matter who is working on your projects?
Interesting thought. In some areas developers are becoming commodity. It's not clear how to interpret the survey data as it's subjective. "Coding as a Hobby" has same numbers 80% for both overall and professional - so it make sense that "80% of our respondents say that they code as a hobby". However, that does not mean only 20% were professional developers.