I wrote on this when it came out and this article link is dated, but it had a different “slant” on things that my first posting:
I note in this article there were comments about standards and not following them, as well as references to the earlier NASA report claiming there was not a software issue.
Now I believe standards have a place, but still need thinking. Some people use them so they don’t have to think about something like testing. Not good. Also, as far as “reports” and investigations go, they need to be subjected to thinking and the “scientific method” in which we question any report to see if the information is incomplete. This is a fact of life for testers (just because your first 100 tests work does not mean that a bug is not there).
So I write a lot these days about skills in testing (building them), using many different approaches to testing, and watching for things like bias. Do Toyota or other companies have bias and need to improve their testing skills?
Probably as bias is a fact of being human and every tester I know can improve their skills.
What can we as tester do about it?
Test more and practice our skill?
Something that is “good enough for government work” fulfills the written specs and standards, and passes the tests for compliance with those standards. This phrase isn’t used to refer to things considered to be of high quality. It is generally used to indicate poor quality.
Yes, “good enough for government work” as a phrase can be an indicator of questionable quality. For me, the problem comes if testers only do verification/compliance checking of things like standards and/or requirements. For those in the regulated environments, I like to say “checking is necessary but not sufficient”, but all too often checking is all that is done and even that checking is of questionable compliance.