Why does the logic of "taxpayer funded research should be free" suddenly change when the publisher is a non profit? If that logic is valid, why is the $500 mill

Why does the logic of "taxpayer funded research should be free" suddenly change when the publisher is a non profit? If that logic is valid, why is the $500 million fine? I should probably have said "similar charges" as I said later. Reducing the margins to nothing can at best reduce the cost by 30% or so, which I doubt would satisfy most open access advocates; given that, the real concern has nothing to do with publishers' profit.