> What I take from the whole article is the author severely underestimates the true cost of software.

I was talking about the "costs of ownership", not the "costs of development". And I'd like to think I understand both.