no, nigel's right, don't remove the supply bypass cap. I can see how you drew the conclusion, but it won't solve the problem, it will just create lots of new ones. you can't avoid treating a DC supply voltage as an RF ground, so instead you have to think of another method for switching your capacitors.
You mentioned that you wanted to avoid open-collector outputs (the ones that require pull-up resistors), however those are probably more suitable for your application... in audioguru's diagram, in the top half, where he has switches depicting the "ideal" setup, you can see that the capacitors would either be connected to ground, or open-circuited. an open-collector output is a pull-down transistor, that either pulls the output to ground or open-circuits it (thus the use of pull-up resistors normally), so it behaves the way you would want. This is a method that was used in a function generator project in EPE magazine, to select the frequency output range of the function generator IC, whose frequency was set with an RC pair. granted, that function generator wasn't designed to operate anywhere near to the hundreds of MHz range...
you will still have the problem of parasitic capacitance of the outputs when they are either on or off, and the non-zero output resistance, but if you're really determined to do this by switching caps with a logic circuit then you're kinda stuck with that and will have to tune your way around it.
You might not be able to find your desired logic chip with OC outputs, but you could always just use some NPN or NMOS discrete transistors (driven by inverted outputs if needed) on normal logic outputs to achieve the same thing.
I still think the whole thing is crazy, but at least this might be a step in the less-wrong direction :lol: