Best place to get answers is in the seaplane forum. There are some rules for setting up a floatplane, but I have seen most of them broken to the extreme, and the planes still fly. The trick is to get a middle-of-the-road setup that will fly, then tweak it from there.

The first consideration is the location of the float step in relation to the cg. I like to have mine in the exact same place as the wheel contact point (for a taildragger)...some folks prefer it farther back than that. Too far back, or too far forward and you will run into handling difficulty on the water at speed..I'll get back to this in a moment...

The second consideration is the angle of the floats relative to the wing. (I have seen wild variations on this theme). The bottom line is that with floatplane resting on a flat surface, with the step and the heel of the float touching the surface, an incidence meter on the wing should give you 7 degrees positive incidence.

Once you have these 2 variables nailed down, you will have a plane that will take off, and fly. (if the floats are the right size).

Now the floats can be considered to have 2 separate sections- the part ahead of the step, and the aft part behind the step.

The only function of the front part of the float is to support the model through boyancy when it is at rest or at very slow speed (taxiing) on the water. Except for this phase, we want the front part of the floats out of the water at all times.
The next speed range (on water) is the moderate range. Here we want to replace the floatation boyancy of the front of the floats with hydrodynamic lift from the aft part of the floats. This is done by holding full up elevator and adding power. At this point, the floats are plowing a lot of water until speed builds up.
As the speed of the plane increases, hydrodynamic lift increases, meaning less of the rear part of the float is required to be in the water to provide the same amount of lift. We ease up on the elevator, which allows the nose of the model to drop a bit, and allows the model to accelerate even more, meaning less and less of the float surace is required to be in the water to support the model (at this point the wings are beginning to supply some lift).
At high speed, the model is going to be running in a fairly level attitude, and only a very small part of the float, just ahead of the step is actually touching - both the nose and heel of the float are clear of the surface of the water...this is what is called being "on the step". As water drag has been greatly reduced, the plane accelerates to flying speed quickly and a gentle rotation (back elevator) will allow the wings to supply the lift required to rise from the surface. (this is where the 7 degree positive incidence fits in...giving you room to raise the nose attitude withough having the back of the floats hit the water again.) The technique is very similar to flying a taildragger from land.

Now back to the step position...if the step is too far back, or too far forward, the plane will be riding on more on the front part of the float at high speed..this can lead to directional instability and waterloops. depending on the model, having the step about 1/2" or so behind the cg seems to work well.

If the model starts porposing on takeoff, abort, taxi back and try again.

Landing is easier....most planes require a bit of power due to the extra weight and drag of the floats. aim to touch down with the back part of the floats first. The drag of the water will rock the model forward, reducing the angle of attack of the wings, reducing lift. If it should bouce, it will settle back quickly. If you touch down with the front part of the floats, it is similar to landing nosewheel first on a trike, and the results will be similar.