About CI in scientific (development) projects:
http://www.nature.com/news/collaborative-software-development-made-easy-1.22729
Quoting:
"In 2015, in an effort to save time and resources, the team took inspiration from the technology industry, automating their testing using a process called 'continuous integration'.
In continuous integration, changes to software code automatically trigger repetitive tasks, such as error-checking. Fundamentally, the process simplifies a task that diligent coders already perform. Programmers usually write lists of tests that they will run periodically to ensure that their code still works, just as Neubert's team do.
But a busy team might forget or lack the time to run them, allowing errors to creep in. Continuous integration automates that process so those checks run whenever a change is proposed, saving team members the time they would spend hunting down an error. A team running genomic analyses could spend more time at the bench, while a group developing climate-prediction software could better refine its models. That said, the resulting peace of mind is only as good as the tests themselves: a poorly designed test can still allow mistakes to pass undetected.
The process is common in the commercial and open-source sectors. A study presented at the 2016 IEEE/ACM International Conference on Automated Software Engineering in Singapore found that about 40% of the 34,544 most-popular open-source projects hosted on the coding collaboration site GitHub used continuous integration in some form."
