High-points testing and Selenium
Submitted by
jenn.
on 2005-05-03 11:25.
The "high-points test" is for touching every feature once and making sure its basic mechanism is intact. This includes submitting and accepting a patch, for instance, but not testing every control and combination of controls, or testing a variety of patch content.
Ideally, we should run a high-points every time any change is tested,
no matter how minor or localized the change. The idea is to prevent
surprises like "That broke module submission?! But it was just a
one-line style change in the course linking form!"As the number and complexity of features has grown, though, even a "quick high-points" is taking longer and longer, to the point where, to my shame, I don't always manage to do it even on non-trivial changes. That's why putting the high-points test into Selenium (the automated browser-based tester we're trying out) is so appealing.
BUT.... (More discussion of test automation to come later, since this entry has already sat around for a few days unpublished.)
