Theoretical way to work around the uncanny valley?

Since there isn't much I can really do to test out this idea, I'm throwing it out there for others to work with.
As the solution to the uncanny valley is to make things more lifelike or far less realistic, but we are not quite advanced enough to make fully believable robots, I had an idea for a temporary solution:
What if intentional and substantial things were introduced to distract from the minor visual flaws? People already are fairly comfortable communicating with a computer, but some of the lifelike robots could be creepy. For example, a visible cable tethering toe robot to a computer would be a noticeable and substantial item that shouldn't make the person less comfortable around the robot (unless they dislike robots and/or computers) but it may just serve as a distraction for the fact that the robot isn't capable of believable facial expressions.
Early robots like the Actroid could be downright frightening, but if some obvious visual cue was added to distract from the minor flaws, then perhaps people could be a bit more comfortable around such robots until facial expressions and human behaviors are perfected?

Very interesting point you raise there but if a person is repulsed by a robot that looks and acts very close, but not exactly as a human, the addition of a tethered cable or 'tail'. might not be enough of a distraction from what is directly in front of them. Perhaps the depth of the uncanny valley will reduce once familiarisation of service and personal robotics increases and everyone owns one, like today's smartphones; imagine showing one of them to someone say, 30 years back - not only would they not know what to do with it (not to mention the lack of GSM infrastructure, but they could still play Angry Birds!), they may even be afraid to use it, despite essentially being a standard phone without a tether.

Also the actual use of the robot is important too - if you knew that this, slightly realistic robot is designed to serve you drinks rather than be a stand-in for a human, you'd treat it more like an appliance and assume some (if minor) unrealistic characteristics.

A final thought is the personality of the robot - if you've watched the Ghost In the Shell series/movies (recommended if you haven't), the androids that form the main protagonists have what they call a 'ghost', which in some cases is a human soul augmented with prosthetics, and in others pure AI. Because of their personality, it's very easy to 'like' the characters. In some scenes they communicate using an encrypted comms channel over the net, where you can hear them speak but their lips don't move and it's not peculiar because they do sound human; even the thinktanks are armoured weapons but their cute childlike voices change the way they're perceived and are very likeable. Aiko currently disturbs me not because of the appearance so much, but the pre-programmed vocal responses - if they were more fluid then I'd be happier to accept her despite being largely motionless.

I personally hope that the former holds true, and familiarisation reduces the gap between the two sides of the valley or makes it shallower (which could well be two unique aspects of the problem), but for now the technology still has some way to go before it is even close to realistic in a full-body sense.

Peter