The Need for Testing











Sponsored Links:

My favorite line in the program, Bruce Sterling, The Hacker Crackdown "-- The things they call "software" is not like anything that human society is used to think.The program is like a machine, & something like mathematics, & something like language& something like thought & art, & information but program is not, in fact, none of the otherthings. The protein quality of program is a major source of its fascination. It also makes program very powerful, very subtle, very unpredictable & very dangerous.Some program is bad & buggy.


Some is "robust," even "bulletproof." The best program is that, Which has been tested by thousands of users under thousands of different conditions in recent years. Itbecome known as "stable". This does NOT mean that the program is flawless, error-free. It usually means there is a lot of mistakes in it but the bugs are well identified & wellunderstood.There is basically no way to ensure that the program is free of defects. Although the program is mathematical in nature, can not "proven" as a mathematical theorem, the program is more like language, with the inherent ambiguities, with different definitions, different assumptions, different levels of meaning may conflict.Program development involves ambiguity, assumptions & faulty human communication.Every change made to a piece of program, each new piece of functionality, every attempt to set a default, introduces the possibility of error. With each error, the risk that the program won't meetincreases its intended purpose.Try reducing that risk.They can use the quality control processes to try to prevent the defects of the soft input, but the only thingthey can do to reduce the number of errors & this is to prove it. Following a series of testing & rectification can identify & solve problems.The evidence also helps us to quantify the risk in an untested piece of program.After modifications were made, a piece of program can be run in a controlled environment& observed behavior. This provides evidence that informs the decision to move to the next phase of the project or to help rectify the problem & finally, in some (dare I say enlightened?) The efforts of program development, testing may actually beused to promote development. Following the statistical models & program development methodssuch as usability testing, program development can move from a focus of artistic activity outside ofstructured discipline.This primer is blatantly in favor of the check. I am, & have been for ten years, a tester. As a tester Ihave suffered limitless indignities at the hands of project managers & development teamsresulted in large amounts of stress for me & my team.There's lots of books written about & for project managers & program developers.
This is probably one of them.

No comments: