What is the rationale for letting DataAspectRatio (etc.) be inaccurate while stretch-to-fit is enabled?

Matlab documentation states that when stretch-to-fill is enabled (as it is by default), then "the axes might not exactly match the data aspect ratio, plot box aspect ratio, and camera-view angle values stored in its DataAspectRatio, PlotBoxAspectRatio, and CameraViewAngle properties." But from Mathworks' perspective...why?? For example, what is the rationale for not having the value of DataAspectRatio track the actual on screen appearance? I know that it's possible, because if I make a plot and then set the PlotBoxAspectRatioMode to 'manual,' then the plot appearance doesn't change, but the values of DataAspectRatio do:

figure;plot(0.1:0.1:pi,sin(0.1:0.1:pi))
daspect
pbaspect manual
daspect

Answer 1

[Evan]

From The Mathworks support:

The stretch-to-fill behavior is not taken into account in daspect when stretch-to-fill is on due to the original design of the graphics system and compatibility. Besides the incompatibility, there is an opportunity to improve the functionality of despect by adding options to set aspect ration [sic] even in stretch-to-fill. The developers are aware of this and may consider it for future releases.