I am currently working on updating regression test suite. I would like to add some simple test at the beginning of the regression suite to make sure it is still behaving correctly and all the basic functionality still works (can perform basic operations and also it is able to connect to the database or other services) before proceed with full regression test suite. These test would not be specialized based on type of changes which were made to the software, but would only cover specific subset of regression tests, which would be the same for every new build. They would cover broad range of functionality. Would these test be considered more as sanity tests or smoke tests? Thank you.
2 Answers
In this case, "sanity test" and "smoke test" are interchangeable. Wikipedia acknowledges a relationship between these two terms as well (also on the "smoke testing (software)" page). Both terms refer to tests that can be run quickly to ensure it's worth performing more extensive, costly tests. Since some organizations discourage using "sanity check" or "sanity test", "smoke test" would be the preferable term.
In my field, Sanity Tests are designed as a checklist which include necessary functions somewhere between 60 ~ 70 test cases. So if the new version passes Sanity Testing, we can start regression test cases.
And smoke tests aren't managed by the QA Team. It is handle by Dev Team.
But I'm not sure of all QA teams are the same as my field.
