Design issues list the test automation problems that can occur when an efficient testware architecture and maintainability are not built in from the very beginning. The table below gives a short list of the issues. Clicking on an issue shows more detail and the patterns needed to solve it.
The Design Issues Mind Map shows an overview of the Design Issues, together with the top level of resolving patterns, both the Most Recommended and other Useful patterns.

Issue
Description
BRITTLE SCRIPTS
Automation scripts have to be reworked for any small change to the Software Under Test (SUT)
CAN'T FIND WHAT I WANT
There is a script or file or dataset but you don't remember what it's called or where to find it.
COMPLEX ENVIRONMENT
The environment where the Software Under Test (SUT) has to run is complex
DATE DEPENDENCY [3]
Tests are dependent on a specific date
GIANT SCRIPTS [3]
Scripts that span thousands of lines
HARD-TO-AUTOMATE
The Software Under Test (SUT) doesn't support automated testing very well or not at all.
HARD-TO-AUTOMATE RESULTS
Preparing the expected results is slow and difficult
INCONSISTENT DATA
The data needed for the automated test cases changes unpredictably
INTERDEPENDENT TEST CASES
Test cases depend on each other, that is they can only be executed in a fixed sequence
LITTER BUG [1]
Tests leave spurious data in databases, history etc. that slows performance or hinders further tests.
LONG SET-UP
Set-up of the initial conditions for the test cases is long and complicated.
MULTIPLE PLATFORMS
The same tests are supposed to run in many different operating systems or browsers
OBSCURE TESTS
Automated tests are very complex and difficult to understand.
REPETITIOUS TESTS
Test cases repeat the same actions on different data.
SORCERER'S APPRENTICE SYNDROME [2]
Automation mimics manual tests without searching for more efficient solutions.
TOO EARLY AUTOMATION
Test automation starts too early on an immature application or on the wrong aspect of an application, and produces only “noise”.
TOOL DEPENDENCY
Test automation is strongly dependent on some special tool.
UNAUTOMATABLE TEST CASES
Existing test cases are “unautomatable”, i.e. difficult if not impossible to automate.

Home
Back to Test Automation Issues
Back to Management Issues
Forward to Execution Issues



[1] Suggested by Thorsten Schönfelder
[2] Suggested by Michael Stahl
[3] Suggested by Dave Martin