Dear Snapper,
(did not look in here for a while).
1. Unless the new Shimano Inter L is completly different that all the others (and I know only about one bicycle dynamo which is completly different), you have a basic misunderstanding: Most bicycle generators are constant current sources. As things are not perfect (and there is no real need to be), they just behave much more as a current source than as a voltage source.
It's output voltage depends on the load resistance, as soon as the minimum speed is reached. That means twice the resistance is twice the voltage and twice the power. Of course, if you need more output voltage, the minimum speed is higher.
Our dynamos here are completly different to batteries, they do not deliver (more or less constant) a voltage like 9V.
2. Saturation does NOT depend on the power delivered. Let's assume the insulation is good for more than we need, you could get easily 30W (and probably much more)out of a SON without any saturation. People have run 220V lightbulbs from a SON for demonstration (at horrible RPM of course). Saturation comes when the current is too high. As constant current devices do not deliver much more than their rated current, this is nothing to worry about. But not all devices are equal. Where have you measured saturation on a bicycle dynamo? I'm very interested in the model and the circumstances? Was it on an Inter L?
3. I have never seen a bicycle dynamo molten down due to overload, although I experimented with them quite a lot (and used them even more). Which brand was it under which load?
4. Alternators are no constant power devices, the are (as stated several times before) constant current sources. The high voltages you mention, came from a resonating circuit, funny enough I was never able to experience them by myself. I know about nobody ever hurt (electrically) by a bicycle dynamo. On powerlines, this is called Ferranti-effect.
5. What you seen in Myra's diagrams ist the voltage behaviour when you load a dynamo with the usual bicycle light. That means, when your load is different, these curves look different. When you put two 6V/3W bulbs in series, the lines will look similar, but the scale will be different. Myra is very knowledgeable and will tell you nothing different than I, probably in a better language .-)
6. When you use a good full way rectifier (either a low drop or a synchronous one with almost no drop) you will get charging voltage for a 6V SLA pretty easy. As the peak voltage is 1.4 times higher, starting starts at pretty low speed (shloud be easy to check). As long as the rectifier is OK, there is nothing to worry about the battery discharging through the dynamo. Beeing cautious is alway good and the use of a fuse is not bad at all (I never use them, BTW).
7. SA dynohub: They were only rated at 300mA, for &V that meant 1.8W, so just in case somebody finds one somewhere, do not be disappointed. And do not disassemble them (a long story). And 4.8V lead acid?
8. A 7.5 is a simple and good solution, but will start to waste energy at pretty low speed.
9. DC/DC converters are available in all sizes and tastes.
10. Again: A 4.5Ah SLA will NOT deliver 4.5 A for an hour, maybe for 40 minutes or less.
After all: I will check the Inter L as soon as I have time if it is really that different as Snapper writes.
In the menawhile: Happy Chrismas
