Saturday, February 11, 2012

Would there have been a formal validation plan for Collins ? (Surely there was ?) Was it ever used?

May 22, 2010 by  
Filed under software development schedule

It’s a bit rich telling the supplier his system doesn’t work, if the formal validation plan was never followed.

< A/ by Jeffrey C/If you don't have a background in systems and software development, this response will give you a quick overview. It is far from complete, however. If you do have a background in this subject, you should be looking at your processes and procedures rather than trying to point the finger at one or more rogue developers.

Programs contain mistakes for lots of reasons. More often than not, problems occur at because the stated requirement is not really how the system needs to work.

The next most common set of errors occur at the interfaces between subsystems, the software and hardware, and between units within the software. One of the most famous of these interface problems was the Mars spacecraft that got lost because one team used SI units but the team they interfaced to used English units.

In real time and embedded work, we often see errors that occur because of the way the software interacts with the hardware. The most difficult to fix are the transient problems that do not reoccur easily. These errors often stay hidden in the code, showing their ugly heads occasionally to bring down a system.

The easiest problems to fix are those in the implementation. We perform unit test to confirm that the code meets the design (no, not code meets the requirements). Note that this testing does verify that the design meets the requirements, only that the code works as designed.

There are several levels of testing that prove the software meets the requirements -- integration testing, hardware/software integration, verification testing. I'm really not going to go into these. Suffice it to say that most mature organizations confirm that each requirement is associated with a test case and that each test case has executed.

The most important testing comes at the very end and often does not get the attention it deserves due to budget and schedule concerns. Validation testing confirms that the requirements are correct, complete, and consistent. This testing runs systems through real-life scenarios to be sure the system performs as desired.

At this point, you probably know a great deal about Collins since you've spent at least 2 years talking about it on this forum. You may be able to find out what the formal validation plan was from the MOD or the contractors. I don't know if this type of information is available to the public, but the adequacy of the plan may help you determine if there was a conspiracy of just a few people who destroyed this VERY large program, or if there were more basic problems due to an immature organization not really knowing how to take on such an involved task.>>

Comments

2 Responses to “Would there have been a formal validation plan for Collins ? (Surely there was ?) Was it ever used?”
  1. jdk2509 says:

    Are you talking about Phil Collins ??

  2. Froggie says:

    Of course but you can never diminish or under estimate human contents, thats why programs are continually changing and upgrading, but if its man made you can always expect a flaw, we dont think the same as a computer but then a computer cant define words by just giving the first and last initial staying the same, but a human can :)

Speak Your Mind

Tell us what you're thinking...
and oh, if you want a pic to show with your comment, go get a gravatar!

You must be logged in to post a comment.

Powered by Yahoo! Answers