Hi Justin,

Thank you for your comment.

This is indeed an important point.

The problem, from my experience, is that it is very difficult to predict the lifetime, the number of executions or the number of users of a program at the time of its development.

How many “temporary” developments actually run for years, simply “because it works”?
How many developments designed quickly “to test” have been pushed to their limits because of an unexpected success?

Productivity is always the first cost item at the time of development, but the predictivity of run costs (themselves depending on the predictivity of the success of a development) being a science that is, to say the least, uncertain, it is advisable, when possible, to be cautious about this priority to productivity.