by eValid » Thu Mar 21, 2013 9:24 am
To "regress" means to change outside the established parameters of success. If you have an application for which your tests evaluate 100 factors, when any one of them changes, then the test is said to have detected a regression (in the application).
In most practical testing situations a single test playback in the simplest form simply drives around on the web application and visits a number of pages. In many cases, just knowing that all of the links and connections are working OK is sufficient.
Along the way, if YOU believe that there is some fact about any of the pages involved then there are simple Validation Commands, available from the eValid GUI, that let you record that you want those facts to be true at some later time. As you do this, for each such validation step you write, you are increasing the chances of the test to detect regression.
The key in this is YOUR knowledge of the application, and YOUR knowledge of what is important to detect. Or not. For example, you probably do NOT want to validate on a date and time value in a page because you know in advance that those are likely to change. You might validate on a month, knowing that you want to be advised, when that validation FAILs, that the month has indeed changed...next month. And, if that test does NOT FAIL next month then you have another regression detected.
Anything that you can see on a page can be validated. And, using the PageMap and the advanced validation commands a lot of things that you can't see can also be validated.
eValid Test Team