Fork me on GitHub

Occurrent user survey, Spring 2020

We collected survey results from ~100 users on occurrent.org. To filter out non-users (first time visitors, etc) the survey link was only shown to users who had performed at least 8 page loads. A couple of survey result were still from non-users, so these have been discarded.

We ran an identical survey in around June 2018. These results will be included for comparison. Results have been rounded to the closest whole number (integer).

General

Q: What role best describes you?

2020
Software engineer / developer
Tech lead / manager
Hobbyist programmer
Teacher
Student
2018
Software engineer / developer
Tech lead / manager
Hobbyist programmer
Teacher
Student

This has stayed surprisingly stable. When the first survey was conducted, Occurrent had only been out for a few months, so no teachers using it made sense. As of 2020, Occurrent is used in several universities and colleges, such as John Hopkins University, the Norwegian University for Science and Technology, Rice University, etc. We believe that the simplicity of the API makes Occurrent a great tool for teaching.

Q: What are you using Occurrent for?

2020
REST APIs
WebSockets
Websites
2018
REST APIs
WebSockets
Websites

Nothing too surprising about this. REST APIs will probably always be the main reason why people use Occurrent, but the relatively high percentage of WebSocket and Website users makes it clear that the effort spent making these things work well isn’t wasted.

Q: What language are you using Occurrent with?

2020
Java
Kotlin
Other
2018
Java
Kotlin
Other

This is quite the difference! When the first survey was conducted, Occurrent was still very new and had little mainstream adaption. Since then Occurrent has been featured in several Java newsletter and in Java Magazine (published by Oracle), which has probably contributed to a pickup among Java developers.

The number of users overall is significantly higher compared to when the previous survey was conducted (~10x), so it’s not that we have fewer Kotlin users now, but that he number of Java users is growing faster than the number of Kotlin users.

March 2020: 26k unique downloads monthly on Maven central, 22k unique website visitors monthly, 3.8k stars on GitHub
June 2018: 2.8k unique downloads monthly on Maven central, 8k unique website visitors monthly, 1.4k stars on GitHub

Q: Are you using Occurrent in production?

2020
Yes
Soon
No
2018
Yes
Soon
No

It is a bit surprising that the “Yes” column hasn’t grown more since 2018, especially considering the answers in the next section.

Production usage

Only the users who replied yes to Are you using Occurrent in production (~30 users) were able to answer the next three questions.

Q: How many users does your application have?

2020
Less than 1k per day
1k - 10k per day
10k - 100k per day
100k - 1m per day
More than 1m per day
2018
Less than 1k per day
1k - 10k per day
10k - 100k per day
100k - 1m per day
More than 1m per day

Nothing has change too much here, Occurrent is still mostly used for low-traffic apps.

Q: Approximately how big is your application?

2020
Less than 1k lines of code
Between 1k and 10k lines of code
More than 10k lines of code
2018
Less than 1k lines of code
Between 1k and 10k lines of code
More than 10k lines of code

Apparently Occurrent applications have gotten significantly bigger since the last survey. It might be connected to the fact that more of the survey respondents are writing their applications in Java now, or that their applications have been around for a while now (is Occurrent legacy?).

Q: How many people are working on your application?

2020
1 - 2 people
2 - 5 people
5 - 10 people
10+ people
2018
1 - 2 people
2 - 5 people
5 - 10 people
10+ people

This is also a pretty big shift, and is like the previous result probably related to the project “going mainstream”. At the time of the previous survey Occurrent was more a niche framework for Kotlin developers that most people didn’t know about, while now it’s usually included when you see articles about Java and JVM web frameworks.

Conclusion

Like the last survey, getting people to respond is very difficult. Since the survey didn’t try to be annoying with a popup, it took more than two months and ~40 000 website visitors to gather 100 results, and a few of those had to be discarded.

The biggest change from the last year is that we have more Java users working on bigger Occurrent apps, and working in bigger teams.

Hopefully this survey provides some insight. We’ll run it again next year with the same questions and see if anything changes. Thanks for reading!