Friday, May 31, 2013

software design paradigm "convention over configuration" applied in software testing

Convention over configuration - Wikipedia, the free encyclopedia

I created a testing framework for one of my projects, and for quite a while I maintained the test cases in a "configuration data structure" (luckily) with rather strict conventions. Let me call that a CDS!
I listed every single test case with all its necessary details there. That was rather cumbersome and did not make me eager to introduce new test cases.
Looking at the fact, that it cost me less than 20 hours to "cast the conventions into code", makes it look rather unclear, why I had hesitated for a couple of months to do so.
Now I am able to drop any "productional data case" into my repository of test cases, and start from there with test driver. That is quiet a relief.

The difference between a raw test case taken from a production system and a (complete) test case executed by my testing framework is "reference data". My testing framework is in a way supportive resp. smart enough by helping in creating "reference data" from the corresponding test data. Certainly a human being must consider the crucial difference between test data and reference data and has to adapt the reference data accordingly.

To be continued.

No comments: