By this morning, all the components had arrived, so I rolled up my sleeves, went to the workbench and built myself a cooker. The result is in my photo album.
By the end of the afternoon, it was ready for trials, and I cooked my first meal on it.
Problems:
The gimballing is not very good. Although the pivots are 6cm above the pan supports, so that they are near the free surface of the liquid in the pan, the centre of gravity of the whole assemblage is still not far enough below the pan, and the cooker is not heavy enough. This shows up in a way that I hadn't anticipated: A pan with a heavy handle, like my Kuhn Rikon pressure cooker, has its CG quite far out from the centre of the pan itself. The cooker will tilt, under this influence, and the pan will slide to one side, increasing the effect. Even my kettle, which is symmetrical and has no handle out to one side, will cause the cooker to tilt if it is not exactly central, even though it can move only 1cm. The only cure would seem to be a large lump of lead bonded to the bottom of the cooker.
The tank cannot be much higher than the built-in tank in the Maxie cooker, or the needle valve cannot close off the flow completely. It cannot be particularly well made. If the tank was higher, there was a fiercer flame than I thought was good for for the burner.
The knob softened in the heat transmitted along the spindle, and would not turn the spindle at the end of cooking. I was glad of the on/off valve near the tank. I will have to find a more heat-resistant knob.
All in all, though, I think the concept is viable. At least I have a cooker that is safe enough, is easy to light and adjust, is fast, is easy to refill and will retain a pan in a rough sea.