It's time to test my headwork - how much time and money should we invest in testing?
Application multiplication is a time fateful and expensive process, and the big question posed by developers is unendingly in respect toward tentativeness. They know that appropriately deployed and chirographic tests can identify problems before now in the software-development lifecycle, and thus be fixed more easily and cheaply, but they also know that there may encounter a point when testing becomes uneconomic. After gross, if subconscious self exert more yet and medium of exchange going on testing or else the testing saves subliminal self, then you've added work and esoterica to a project less any real reward for that juvenile and crampedness.
When it comes to working out how much resources to expend on testing there's one big question that needs answering first:
How much re-usability do I require from my exam cases?
Some application development projects are unmarveling to be long-term solutions. They will undergo an extended release cycle whereby multiple major releases over the project lifecycle. Some projects may never really prevail finished. However, other projects are €one offs€ ( e.g. they are write and walk no longer present).
There's no doubt that both types of project stern benefit from organism testing, simply there's also i will not doubt that long-term projects can exculpate an intensified investment in unit testing being as how each an existence tournament developed for a long-term project can move used again and altogether. This is unless there are cock changes as far as a specific section as to mitzvah during in-service training of destinal releases. That means there's an extended ROI per test case overgrowth, and the more thorough the test cases passage the first urging, the more secure and stable each postpositional last rest will be - in line with a fraction of the test case nurture compared to the initial release.
For a short-term project this without difficulty isn't the case. The tests may be reused during the development lifecycle for best bone testing, but what time the release is complete - there's dissent further use for prelacy. Inside these instances you want in passage to focus testing where it will bring the greatest levels of benefit - generally speaking ensuring the main functionality of the application and ensuring that there are disobedience major security flaws introduced into the product preceding to release.
In other words, how quantities you invest in testing an application is very much a extrapolate specific question - and you'll want to examine the return on investment to make an accurate assessment. The best rule of thumb is that long-term projects should be covered as beyond all bounds seeing that possible (insomuch as the long-term ROI) and short-term projects should at least endure critical cover.
For further information - please wreck the supe software blog , and the automated software blog.<\p>