November 25th, 2019

(Though, based on the amount of ink spilled, it certainly seems like 4/5 devs are using “framework of the month”.)

(Though, based on the amount of ink spilled, it certainly seems like 4/5 devs are using “framework of the month”.) adactio.com/notes/16167

Ink spilled?

The source for my data is an MDN developer survey. It definitely has survivor bias though, so we should probably survey folks who left web dev to go to android or ios development.

I’ve looked at several data sources: This year in JavaScript The state of JS Stack overflow survey The state of CSS MDN survey MSFT forms surveys HTTP Archive NPM install data Chrome data Each has biases, but in total they paint a picture.

It’s important to be clear, my stat was about developers, not sites. For sites you need other data sources… e.g. twitter.com/stubbornella/s…

I knew a lot of freelance designers and developers before I moved to SF who would never hear about either survey.

To that end, important to acknowledge the bias in HTTP Archive… it’s logged out views of only top level domains. (so it’s much more likely to hit marketing sites and login screens)

That’s why I look at so many different data sources… including installs and web usage. Also look at engagement weighted metrics like page views and time spent. twitter.com/stubbornella/s…

I don’t really think that impacts what I am getting at. Some of the people I kept in contact with still don’t use npm. I also personally missed more than half of those surveys this year. Metrics are important, but they can erase people.

We also have a lot of conversations (like these). It’s more qualitative but it counts too!

sites != devs?

Yes. If you look at sites, it’s all still jQuery and backbone. 😅