"Steam Rockets: Jim Powell at Brookhaven calculated it would take about 800 Celsius to get Isp of ~ 215 seconds specific impulse. He used nearly perfect nozzle expander estimate. In practice one can get 95% perfect, but I am not sure what the 95% means: energy efficiency or Isp efficiency."

I assume neither; rather, 95% efficient in terms of a nozzle shape/size.

Though Zuppero does talk a lot about energy efficiency- he says fuel efficiency is a wrong focus:
"If we run a larger amount of propellant through the heat exchanger, the propellant will keep the heat exchanger cool.and the propellant gas will be cooler. The speed of molecules coming out the nozzle will be slower.

If we run a smaller amount of propellant through the heat exchanger, the heat exchanger will heat up and the propellant gas will be hotter. The speed of molecules coming out the nozzle will be faster.

Which way should we go? Run lots of propellant through the heat exchanger and get cooler gas, or run the least we can get away with and get the highest exhaust velocity?

Rocket scientists demand that we use the least rocket propellant we can, and that we achieve the highest exhaust velocity we can. "
And:
" “Highest velocity” is the rocket scientist’s incorrect answer. "
http://www.neofuel.com/optimum/

My summary of what Zuppero is talking about, he says rocket scientist, try to conserve fuel [or mass of rocket] by trying to make exhaust the fastest- throw less, faster. Zuppero thinks that in space environment one should be more interested in using less energy- throw more, slower.

Now, this could make sense if you can afford to waste propellent and desire to use at least amount energy to gain the most thrust [and particularly if you using nuclear energy as source of energy].

My main disagreement with this idea in general, is that the water propellent could be relatively expensive and the cost of energy in space is not necessarily so high that emphasis should focused solely on using the least energy.
If we had lakes on the Moon- water in vast easily accessible quantities I might think about it differently. E.g. with Europa it might be workable.

Zuppero gives this "formula" and rule of thumb:
"This means that the best propellant exhaust velocity is about 2/3 of the mission delta-V"
And he means "best propellant exhaust velocity" in terms of using at least energy. I don't know about his guesses or his math, but do know LH&LOX exhaust velocity is about 2/3 of delta-v required to reach Earth Orbit.

I'm not willing to accept this as some "truth" that always "works" but maybe it's an acceptable rule thumb. It might allow one look comparatively at something. For example, on the Moon 2.5 km/sec delta-v will get you to L-1. So, according his rule of thumb, an exhaust speed of 1666 meters/sec [170 isp] will give about same rocket acceleration/characteristic as rocket with 450 isp launching from Earth. [But of course launching from Earth sucks- it's hard to get stuff off the Earth.] So, a rocket the same size of Saturn V on the Moon with a 170 isp would put over 100 tons in lunar orbit. In other words, a 6 million lb [3000 ton] rocket with around 170 isp could according the rule of thumb get 100 tons to lunar orbit. Whereas if you using 450 Isp engine it would take about 400,000 lb [200 ton] rocket to get over 100 tons to lunar orbit. Now, you probably could make a rocket that uses compressed air get off the Moon [called a cold gas thruster which has generally less than 100 isp {superduper high pressure tanks or other things could improve the ISP]- btw, you could not do so from Earth- though not sure about Mars. Probably someone has already thought of such an idea if it was possible [vaguely practical], but obviously Mars does have an atmosphere and with Mars changing temperature one could get compressed Martian air using little energy*

Next:
"Powel had tested some steam reactor elements at 1500 Kelvin. He believed we should say "235 seconds" so he could use his particle bed reactor pellets he tested. "

According to Zuppero:
Corrosion and oxidation of structural materials generally limit service life in high temperature steam. Although only limited long-term test data are available, one study [Berry] ranks several nickel-based alloys on oxidation and corrosion rates after 4000-hour exposure to 1000 K steam."
http://www.neofuel.com/moonicerocket/

This seems somewhat reasonable. But the 1500 K is pushing it. The highest steam generator on earth is somewhere around 1500 F [about 1000 K].
"Because the majority of the U.S. electricity production and much of its industrial heat depends upon steam, the DOE has launched this development to make a step change in technology to 1500 deg F steam. The previous 100 years of industrial and utility developments have only averaged about 7 deg F per year."
http://www.ms.ornl.gov/programs/energyeff/ats/highperf.htm

*Frozen CO2 on Mars could be used as an energy source- maybe not having the energy density to make a good choice for a rocket, but it could be used to run vehicles, electric generators and/or powertools.
And of course liquid water in space is also something with potential energy- it takes a fair amount energy to turn ice into liquid water and gives up energy turning back into solid.