I dislike the use of the "robot" for anything that doesn't have autonomous movement (those were "RC Wars", not "Robot Wars"---although I, for one, would have delighted, at a safe distance, in someone strapping a circular saw to something autonomous). Nor is watering a single indoor potted plant the same thing as "gardening." An actual Gardening Robot would have been very interesting (I envisioned something trundling around with a spade and actuators for pulling weeds when I saw the title).
I actually love the Pi---one is now my primary computer---but it seems to have created a niche for "let's add a website and database to my really trivial control systems project" that I'm not sure really advances much of anything.
I see your point, but I think that complexity is key (and I think that definition 3 is far too over-broad, unless you require minimal complexity of the mechanism or pay close attention to the plurality of "controls"). The problem with categorizing anything with a sensor and actuator as the proper study of robotics means that very little is excluded. We wouldn't be impressed if a car salesman claimed that the automated dash-light dimmer is a "robot" (even if he had a well-researched semantic argument), and I don't think the term is applicable in this case.
I actually love the Pi---one is now my primary computer---but it seems to have created a niche for "let's add a website and database to my really trivial control systems project" that I'm not sure really advances much of anything.