Chapter 15 mentions that critical systems require advanced testing due to the costs of failure and validation of dependability attributes (you may have to prove that you tested it and it works). Personally I find these two things to be thoughts that haunt me as I develop regardless, and couldn't imagine releasing a critical system without rigorous testing. Rockets shouldn't fail in launch or orbit due to a simple, basic test that was never run due to laziness or oversight.
Static analysis comes across as the basic overlooking and viewing that happens as you write the code, as well as the simple code checkers and peer review. A necessary first step, one that covers a lot of basic and disastrous conditions but still overall not too complicated.
Reliability testing is something that's important as it seems like it can uncover some real problems going on under the hood. Things go wrong for a reason - if the reliability of your system is really low - a simple change or stomping of a subtle bug may really improve it and lead to a better user experience.
Security testing is definitely more important now than ever before. Though I'm not experienced at all in the area of security, I recognize it's importance and I'm interested in it. I'm personally up in the air which is more important to me personally - reliability or security. On one hand having a reliable system that's vulnerable means that those who get unauthorized access have correct in formation - on the other if a system is unreliable, can it really be secured if the security is part of the same design?
Process assurance seems to reaffirm something that we've been taught in general in life - products made of better materials generally make better products.
No comments:
Post a Comment