In my opinion, a really big part of testing a web site is functional testing (also called black box testing).
Generally a functional test is based on the program specification. There is no difference concerning testing a web site. The web page specification is the basis to develop an functional system test. Maybe it's called "Release Documentation" or "System Requirements Specification" or you have all of those documents (this would be redundancy, or not? [img]/images/graemlins/smirk.gif[/img] ). Nevertheless, those documents are the input to develop an functional test.
The object of black box testing, is to test the specified functionalities as comprehensive as possible but without redundancy.
In software developing, there are different methods to determine test cases.
<ul type="square"> [*]equivalence partitioning [*]boundary value analysis [*]error guessing [*]special values testing (--> developing error sensitive test cases from experience) [/list]
1.) What do you think about this approach?
2.) Which methods do experts (YOU?) use in practice?
3.) Do you know and use other methods?
4.) When you build test cases, how do you proceed?