Not silly at all. The bacterial decomposition and chemical corrosion would obviously make a difference to the strength of a rivet or plate as a whole component, but they wouldn't measure it that way.
The corrosive effects generally wouldn't penetrate through the full thickness of the component so they would cut out a bar of measured dimensions from the undamaged inner iron or steel and measure the tensile strength of that, scaling it to the standard units of newtons/square metre. Then they would use that value for the tensile strength of the material to calculate the tensile strength of the original, uncorroded components.