I've had some dealings with some exotic gas blends before. To make it simple, the lighter the fuel and higher the requirement, the faster the floats should work. That's a bit of oversimplification, but it does work like that from what I've seen.
As for ethanol, I'm all for it. I had a very rare chance to watch a guy tune a ProEFI standalone ECU that uses a Ford ECO-Boost fuel sensor in the tank to measure ethanol and gas levels. He did a full tune on Ethanol and a full tune on Gas. The ECU would account for different levels by merging the 2 maps together to get the AFRs and ignition right. After seeing just how differently the engine acted on both fuels and hearing some of his theories, I learned quite a bit.
I don't think that ethanol gas is bad at all. It has higher anti-detonation resistance than gas, but requires much lower AFRs than gas. When you have a 10% ethanol mix, it equates to running at 96% of the original fuel volume required for a proper AFR. That isn't enough to have to set floats, but it is enough to have to re-jet. That's about the same as running 25:1 gas/oil. Actually, running 10% ethanol gas makes an engine run leaner than switching from 40:1 to 20:1. I'll have to start mentioning that in jetting recommendations.
Some people say that if you leave mixed gas sitting that has ethanol in it that the oil will separate after a few days. I tried it and the only thing I could notice is that after a week, the top and bottom halves were a slightly different shade in the glass container. If anyone has ever opened their gas tank while putting along, you'll see why keeping the gas mixed isn't a concern when it doesn't separate that much.
I'll get back on topic now. As for floats, I've never never had to re-set them on anything that I've built, but I don't build 70hp machines either.
Now if you get into specific weights and their relation to the octane rating (which isn't linear, but suggestive), it does make the mind boggle that some tuning company, who's name I can't recall, did a test on some 93 pump gas and it came back at 108 RON! They had a problem with tuning cars and customers coming back with their cars in safe mode because the knock sensors went haywire and threw codes. In that scenario, it does make you wonder.
IMO, specific weights don't really come into play unless you're already at the point of out-flowing your system as a whole. By the time you get to that point, a difference in specific weight of .75 and .72 will be the concern of only those who want to stay with the current setup for whatever reason.
Within the scope of what I currently deal with, it all boils down to the super-simple technique of plug chopping.
