There isn't any kind of really large astronomical instrument which could be operated more simply & with greater flexibility on the Moon, especially if there was a small, habitable zone nearby to allow easy upgrading of the technology.

Because the Moon has only an exosphere, any form of optical (optical used here in its widest sense, which includes IR; xray; etc) telescope could operate continuously (ie, the 14 days of sunlight are irrelevant) if the scope was properly baffled (down a crater, to use only the most obvious eg; & targetted to avoid internal reflections) & aligned; interferometry is far easier because you have a stable platform from which to work; & the far side has a radio silence which we on Earth are rapidly destroying. The polar regions provide more traditional terrestrial-style days (ie, you don't have the 14 days sunlight/14 days darkness of the equatorial lattitudes), which makes power options more flexible; & communication is obviously a lot easier.

Even Ed's assessment of the surface as a platform for NEO impactor research is basically accurate; although (as always) he massively overstates both the degree & immediacy of the threat in order to prop up his argument for this expenditure. There is no immediate NEO threat which would justify such a level of expenditure on the Moon (or elsewhere; although a small increase in terrestrial observation capacity supplemented by a small space telescope capable of both Aten detection & infrared remote sensing is certainly supportable); but in a political environment where human space was being reconsidered, NEO research would be a valuable sidebar.

More generally: the question remains whether this potent scientific platform is really worth the expenditure required in setting it up using current & near term technologies; but if the political will is that it is, the science case on the Luna/Mars issue is unquestionably in the direction of Luna (in cf to the Moon, the Mars option is little more than an Apollo-styled jolly jaunt with little sustainable scientific potential)... but watch out for dark horses like 2002 AA29. Asteroids in near terrestrial orbit which require low delta vs (something which can be easily approximated by checking the object's Earth flyby speed... about 6kmps in the case of thsi object) & have short mission times allow human missions pretty much with off the shelf technology at budgets (if Russian heavy lift is used as is) competitive with the cost overruns of pork nonsense like ISS.

We don't have too many of these objects checked for mission delta vs at present; but research is continuing: so in a few years time, the idea of launching a MIR/Skylab-class manned platform on a slow flyby or even a true orbital mission to one of these objects is going to look very cost effective indeed. No lander technology is required (you can literally space walk from ship to surface); communication is a breeze (2002 AA29 never gets further than about .2AU from Earth) with no power issues to limit your flexibility (large solar panels are fine in terrestrial space). We'd have to modify Soyuz technology for the reentry capsule (return speeds would be higher than that achieved by LEO reentry) because only the Russians & Chinese are working on real reentry technology currently; but with Proton as you base lifter & modified Russian rocket stages (possibly including Proton again; although Soyuz is also possible) as disposable escape thrusters (probably with the assistance of a Lunar/Earth flyby to achieve plane changes... a fairly simple manouevre using multiple burns out from LEO via the WSB; but which has never been seriously considered for human flight), the only real technological problem which would have to be addressed here is the ability to store your return fuel for the several months of the mission before firing... & that in itself could be tested using OTS technologies by simply placing a suitably modified - & fueled - Proton stage in LEO & doing a delayed fire.

All up: this is something which could be done in the same order of cost magnitude as our present shuttle expenditure & without massive technology development. Much more appealing in the current political environment; & who knows? Cost effective human space technology like this might even make returns to the Moon & jaunts to Mars more feasible in the long term.....

Robert