If you follow alternative energy news (where I define "alternative" as anything not derived ultimately from fossil fuels), you get deluged with announcements about promising new technologies, each of which seems to be poised on the brink of ending all our problems.

Most of this is hooey. There's a lot of money flowing into energy technology these days--for good reason--and so there's a lot of incentive to hype any new invention which might somehow conceivably be pressed into service for energy production.

Since most of the stuff you read about will never produce a single watt of power, it's important to keep some perspective on exciting new developments. Here's my list of questions you should ask about any new energy technology:

Has the technology been demonstrated outside the lab?
A lot of hype gets expended on technologies which exist only in the lab, and sometimes only on paper. There's a huge gap between a lab experiment showing that geewhizium-doped buckyballs can extract energy from cellulose and actually throwing your grass clippings into your household Mr. Power box and getting electricity out. Most promising technologies never make it out of the lab environment for a variety of reasons: poor operating reliability, sensitivity to common environmental conditions, short operating life, and so on. If a technology has actually been shown to work in real-world operating conditions, that's a huge step forward; otherwise, it may be interesting, but probably irrelevant.

Does the technology produce net positive energy over its lifetime?
Surprisingly, many of the technologies promoted as future energy sources don't actually produce energy by the time you subtract all the inputs. Biofuels, in particular, are controversial in this respect: some calculations show that corn ethanol actually contains less energy than it takes to produce by the time you add up the energy required to plant, harvest, transport, ferment, and especially distill the stuff. The fact that the calculation is even close shows that this is probably not a promising source of energy. Hydrogen fusion is a great example of an energy source which is easy to demonstrate, but has yet to demonstrate net positive energy output.

For a long time, photovoltaic solar cells had the same problem, in that it took more energy to produce a solar cell than the cell could reasonably produce in its lifetime--though manufacturing efficiency for solar cells has long since improved to the point where this is no longer the case.

Other technologies, like wind, hydroelectric, biomass (i.e. burning wood), and nuclear are unambiguously energy-positive.

Can the technology scale?
Humans consume a staggering amount of energy, around 5x10^20 joules per year, about five-sixths of which comes from burning fossil fuels. That's approximately the equivalent of 140 trillion watts of installed photovoltaic capacity. So any technology which is going to eliminate our dependence on fossil fuels has to be able to get big. Really big.

Our global economy has an astonishing capacity to build manufacturing capacity in a hurry: just look at how much infrastructure we''ve built over the past hundred years dedicated to things like refining petroleum, building cars, and processing foodstuffs. So I'm not worried so much about being able to manufacture energy infrastructure as long as the raw materials are plentiful. The kinds of things which can prevent a technology from scaling up are:

Limited resource: Some energy resources, like hydroelectric, geothermal, tides, and (to a lesser extent) wind only occur in a useful form in particular times and places; and while these can be very useful resources, they can't provide energy everywhere its needed. In some cases, the total global resource may be insufficient to replace more than a small fraction of our fossil fuel usage. Waste-to-energy schemes are inherently limited by the amount of trash available.

Limited materials: I've seen some proposed technologies which rely on exotic materials (platinum catalysts, etc.) which may not exist in enough abundance on the surface of the Earth to meet our energy needs. Fortunately, most alternative energy ideas use fairly commonplace raw materials, though there is some question about the availability of uranium for nuclear power (but see this rebuttal).

Limited area: The earth has a finite amount of area, and the space available for power generation for certain applications can be limited by the application (for example, the roof area of a car limits the amount of solar energy it can collect). For biofuels, which are essentially solar-to-liquid fuel, some calculations suggest that the current technology would require more arable land than currently exists in order to replace our fuel needs, though this could change with higher yielding crops and different growing techniques.

Environmental impact: Most alternative energy sources are fairly clean, but some present real problems. Nuclear, in particular, poses some tough disposal issues which haven't yet been solved even for our current level of production. It's not clear how we would deal with the waste produced if we ramped up to using nuclear power for the bulk of our energy needs.

Technologies which can't scale to the 10^20 Joule/year range are inherently destined to be niche energy sources. In my view, the only alternative energy source which has been proven to be scalable to global proportions is photovoltaics: with current technology, the entire global energy demand could be met by covering an area smaller than the U.S. desert southwest in solar cells. Of course, in practice the solar collectors would be spread out across the globe, and a large fraction (perhaps all) of the required collector area could be on rooftops and other currently unusable areas.