It’s that time again
- Fit for Developing Software: Framework for Integrated Tests by Ward Cunningham
- Cory Foy’s blog
- David Chelimsky has created a new sourceforge project ‘.NET FitServer with FIT and FitLibrary’ for the FitNesse.NET code and Mike Stockdale added the FitLibrary.NET code. It’s in the Subversion repository at https://svn.sourceforge.net/svnroot/fitnessedotnet/trunk
- How to include FitNesse pages into CVS
- Isn’t the point of automating tests not to rein in a chaotic system, but to ensure that with a reasonable amount of certainty, the system is behaving as expected.
- FIT is best when it’s part of the design process. RetroFITting will probably be more expensive than it would be worth. RetroFITting is a great term ðŸ™‚
- The biggest advantage of automated testing is the fact that only with automatic tests it’s possible to do refactoring in a pragmatic manner
- Using automation to get you to a place where you can start doing some manual tests is a terrific thing
- Another Conference – Great Lakes Software Excellence Conference
- If you find value in applying SQA/PPQA services to the ML2 process areas, they why WOULDN’T you want to derive the same value by applying it to ALL of your documented processes?” – Mark Paulk, chief architect of the CMM for Software
- PPQA should not necessarily touch the developers however most of the processes that will be checked by PPQA should touch them
- A paper which ‘…suggested a testing approach for determining the number of TC within the budget available to the Project Manager. The point is not only to determine an optimal number of TCs from the technical viewpoint, but considering also their cost, in order to decide which should be run anyway and which eventually in a further moment’
- On the number of Test Cases to produce:
Take a look at the Measurement and Analysis process area, especially specific goal 1; It’s nothing more than the Vic Basili’s Goal-Question-Metric paradigm. Figure out what you want to know (e.g., how many test cases must be written for a particular application), what meaningful measures may be predictors of the identified information need (e.g., estimated number of web pages, estimated number of fields per web page, average complexity, etc. etc.) and construct a “best guess” predication model.
Try using the model, determine what is leading to the model’s wildly wrong
predictions, and adjust it over time. Continue to do so until:
- It’s good enough for use on real projects
- You decide that it’s more effort than it’s worth
and they try something else.