The postal scale idea works well, too, for smaller cats. Someone else on another forum said they were using one, too, and happy with it.

One problem with people scales is the accuracy. The resolution, i.e. the smallest increments in the readout, really isn't a representation of how accurate the scale is, in terms of giving the correct weight. Just because a digital people scale has a readout with 0.1 pound increments doesn't really mean it's accurate to 0.1 pound. Researching this online last night I found plenty of posts where people said they stepped on one scale and then stepped on another scale and the readout varied by five pounds. I couldn't find any good data on the accuracy, though, because the scale specs present readout accuracy as being the same as resolution, which is an improper mix of terminology.

I think as a general rule, the more expensive the scale, the more accurate its sensors and its readout is. If the accuracy (or maximum error) of a scale is a percentage of the maximum weight that it can measure (which is the usual measure of accuracy and is represented as variance) then a scale that can weight up to 300 pounds, as the usual people scale can, and which has a variance (or maximum error) of, let's say, one tenth of one percent (which is pretty good -- and is what the expensive medical scales are), will give a weight accurate to within plus or minus 0.3 pounds (or 4.8 ounces), which is plenty accurate for weighing people. Whereas a scale that can measure up to ten pounds, as the postal scale can, and has a variance of 0.1 percent, can give a weight to plus or minus .01 pounds (or 0.16 ounces) which is plenty accurate for weighing mail. But the people scale's variance of plus/minus almost five ounces, or ten ounces from one extreme to the other, represents a huge error in weighing a ten pound cat.