Consider the following art. The V is centred horizontally but needs to be enlarged. Currently, using the top handle (and shift) to enlarge from the centre-top distorts the image. But if a corner handle is used, the operation unnecessarily becomes two-steps instead of one (enlarge, then re-centre).
Like-wise if the option modiefier is held. Because the V is alinged to the bottom, enlarging from the centre will once again require two steps (enlarge, re-position).
Transform handles should honour the constrain modifier because when the user implicitly tells Photoshop to constrain proportions (or un-constrain depending on what you want), handles should honour this user instruction - right?
That way, we can perform operations in one step. Illustrator/InDesign already work this way as did Photoshop once upon a time - why create multiple steps?
Consistency within apps and across apps is huge so hopefully this no-brainer can get sorted quickly!