Thanks.
So it's not economical, not reliable and not that useful.
Apart from that, it's a great idea!
I'm a bit surprised that it's so unattractive and difficult to make worthwhile. You're constantly moving known mass into and out of the a/c during refuelling and engine operation, giving a great opportunity for auto-calibration. Likewise, slopes are easy to detect and factor into the sums.
But I guess that the number of expensive accidents and fuel inefficiencies caused by weight discrepancies are too low to justify the expense of yet another system...
(This may change. I was talking to a chip maker about embedded supercomputers, where you effectively have a real-time physical model of your vehicle running on a few teraflops of cheap silicon inside the vehicle itself. As things change over time, the model spots non-optimal operations, models 'what-if' experiments on the fly and recommends optimising operational changes. The chip guy was of the opinion that this could compensate far better for things like control surface failures than pilots ever could. Well, maybe. But any system using such ideas would have to be very well instrumented, I'm sure to the extent that the weight and distribution would be accurately known at all times).
R