What Can GitHub Tell Us About the HDL Industry? Part 4

In part 3 of our GitHub study we presented the number of users on GitHub doing their VHDL verification with one or more of the analyzed frameworks: VUnit, OSVVM, UVVM, UVM, and cocotb. The results, especially that for the professional users, came as a bit of a surprise which lead to interesting discussions in the comments of the post. Can a study of professional users on GitHub really say something about professional practices behind company walls? Especially when most professionals on GitHub don't share their professional work but rather private projects.

The first part of this is to accurately establish whether a user on GitHub is a professional or not when the public profile doesn't share such information. This was a bit of manual work but there are many other sources to that information: copyright notices, Git logs, LinkedIn, Google etc. In the end, only a small fraction of the users couldn't be identified.

Second, this study do not expect, nor assume, that the work published by professionals on GitHub is the work produced at their companies. What we study is their practices. Those practices can be the same as the ones used at work or they may choose differently when they can decide on their own. Either way would lead to interesting conclusions. So which one is it? To figure out that we need to analyze where this study is statistically consistent with the Wilson study and where they are significantly different. This is the main theme of the next post but also this week's findings are touching on the concepts of consistency and significant differences.

This week we're looking for anomalies in the data with focus on regional differences. As it turns out only UVM has an even global adoption that is statistically consistent with the global distribution of VHDL users. The other frameworks have global presence but are significantly under and/or over represented in some regions.

For more information jump directly to the anomaly section of the study or read the full story from the beginning.

The code used to derive these facts are part of an open science project. Everything can be reviewed and the results can be repeated. We encourage contributions and suggestions on other interesting facts that we should derive.