If there are problems in your safety-critical software or software release procedures, was your “test section”
“not supposed” to find errors ?If there are problems in your safety-critical software or software release procedures, was your "test section"
Terrific insight
Thnaks.
Sponsored Links:
Random Posts:
Problem arises if the upgrade does not eventuate.
To am,
To joan,
To Epistemologyfan,
For example if the system released is scheduled to be “upgraded”, the test section may be “not supposed to find errors” in the provisional system – intent being that an upgrade, which will fix any problems, is about to commence.
Article Local Url : If there are problems in your safety-critical software or software release procedures, was your "test section"
Comment by Al M
The time to get software done is getting shorter and shorter, and resources made availalbe to test it also smaller and smaller.
The push is really out to get new software implemented as fast as possible, whether it works or not, then fix problems as they are stumbled over.
There are lots of automated testing facilities available at reasonable cost, but very few companies invest in them, for pretty much the same reason very few companies do computer audits or security audits. It is because top executives do not really believe in murphy’s law or investing a modicum of insurance against murphy’s laws.
You posted this question in philosophy … I think it would get better answers in a category of computing or business management.
Comment by joan
This is possible. Was that what happened to my computer? Well it’s fixed now and works great but for the first couple of weeks it sucked. I’ve seen this done with training people also. You do not show them ‘A’ because a will be changed to ‘B’ so soon that learning about ‘A’ will only be confusing in the long run.
Comment by epistemologyfan
Philosophy can comment on any subject, but it may not answer particular questions. I will do the best I can.
I am told that no testing can discover all possible errors; of course, it depends on the particular subject. In computer programming, it remains true. As software becomes more complex, and more people are involved, the matters that can go wrong increase exponentially. Of course, the more skilled the testing group, the less likely of mistakes escaping attention. In a way, error checking is always unpredictable. Mistakes cannot be routine; if they were, checkers would tell the programmers know, so the error would disappear. Such description is the theory.
In practice, business expects to spend in testing only a fraction of the money that was used in the creation of the software, which could be insufficient. The testing group can only detect the more obvious mistakes. Whether they should have uncovered an specific mistake depends too much on the facts of the ground.