The latest expert opinions, articles, and guides for the Java professional.
It’s not often that a JRebel customer employes the scientific method to ascertain the actual time savings enjoyed from avoiding the annoying restart/redeploy process in Java development. Usually, once our users install JRebel and discover the joy of coding in Java without constant restarts and interruptions, they express their pleasure quietly and among their team.
That is, however, until Networked Assets, a German company that counts Atlassian, Oracle, VMWare and IBM among it’s solution partners, decided to see exactly what kind of productivity gain they could get from using JRebel to speed up Java development time.
The results were jaw-dropping. Networked Assets committed to a magnificent scientific study on “Team Velocity” that yielded an astounding 40% gain in development productivity. This was taken from the blog post on the results:
The conducted evaluation shows significant improvement in the Team Velocity once the JRebel tool has been installed. Namely, the Velocity was increased by 40.6%, and according to t-test it was a statistically significant difference. Nevertheless, the obtained results are by no means representative. Only one project and one team was under investigation, and because of that, the external validity is limited. Therefore, it is not possible to assess whether there are some special conditions or circumstances that must be met in order to obtain similar results. Nevertheless, it is also possible that in other project the improvement would be even better.
This is quite a bit above our own averages based on our recent survey with thousands of Java developers, in which we found that 17.5% of development time is lost to application server restarts – which is just over 10 minutes of every hour.
I suppose only congratulations are in order – well done to the Networked Assets team in achieving one of the highest productivity gains we’ve seen yet among JRebel users!
Leave a comment