I thought his statement that all requirements should have a name assigned to them was very good as well.

> " if a product is reaching the end of a production line with a high acceptance rate, there is no need for in-process testing."

> This is really interesting, probably controversial, but worth a lot of contemplation and I wonder how this could apply to software.

Depends on the field. In windows the user will test it with no repercursion to the developer. In other industries every bug which is discovered by customer is expensive.

>This is really interesting, probably controversial, but worth a lot of contemplation and I wonder how this could apply to software.

I was wondering that too. Is it the same as deleting unit tests?

It's like removing bounds checking (a runtime safety feature).

A manufacturing line is more analogous to software during use than it is to software during creation.

I think its kind of like when in Clojure, you only activate validation on the boundaries of your program in production, and disable the validation in the internals of the program.

Because a factory would still test its input, or the input would have been tested by the supplier.

I've noticed that my unit tests are really useful to work out the bugs, but then they might get stale.

Once something is tested and worked out, it's like they become a little redundant.

In software of course re-iterating the tests is maybe cost free, whereas in physical production 'unit tests' introduce maybe non-negligible costs so the 'advantage' in not having to repeat them possibly makes a difference.