Stephen Hawking says that humanity won't survive the next thousand years unless we colonize space. I think that Hawking is an optimist.

We've seen a certain amount of worry about smallpox, anthrax, and various other bioweapons since 9/11. At the moment, and over the short term - say the next five or ten years - these worries, while not without basis, are probably exaggerated. At present there aren't any really satisfactory biological weapons. Anthrax is scary, but not capable of wiping out large (that is, crippling) numbers of people. Smallpox, though a very dangerous threat, is hard to come by and easy to vaccinate against, and the populations whose members are the most likely to employ it as a weapon (say, impoverished Islamic countries) are also those most vulnerable to it if, as is almost inevitable, it gets out of hand once used.

That will change, though. Already there are troubling indications that far more dangerous biological weapons are on the horizon, and the technology needed to develop them is steadily becoming cheaper and more available.

That's not all bad - the spread of such technology will make defenses and antidotes easier to come up with, too. But over the long term, by which I mean the next century, not the next millennium, disaster may hold the edge over prevention: A nasty biological agent only has to get away once to devastate humanity, no matter how many times other such agents are contained previously.

Nor is biological warfare the only thing we have to fear. Nuclear weapons are spreading, and there are a number of ways to modify nuclear weapons so as to produce lethal levels of fallout around the globe with surprisingly small numbers of the devices. That's not yet a serious threat, but it will become so within a couple of decades.

More talked-about, though probably less of a threat in coming decades, is nanotechnology. Biological weapons are likely to exceed nanotech as a threat for some time, but not forever. Again, within this century it will be a danger.

Want farther-out scenarios? Private companies are already launching asteroid rendezvous missions. A mission to divert a substantial asteroid to strike Earth is probably only an order of magnitude more difficult - within the resources, in the not too distant future, of small nations and death-obsessed terror groups, or perhaps Luddites who believe that smashing humanity back to the Neolithic would be a wonderful thing.

No matter. Readers of this column are no doubt sophisticated enough to come up with their own apocalyptic scenarios. The real question is, what are we going to do about it?

In the short term, prevention and defense strategies make sense. But such strategies take you only so far. As Robert Heinlein once said, the Earth is too fragile a basket to hold all of our eggs. We need to diversify, to create more baskets. Colonies on the Moon, on Mars, in orbit, perhaps on asteroids and beyond, would disperse humanity beyond the risk of just about any single catastrophe besides a solar explosion - and type G stars such as the sun don't do that much.

Interestingly, spreading human settlement to outer space is already official United States policy. Congress declared it such in the 1988 Space Settlements Act (Congress declared "the extension of human life beyond Earth's atmosphere, leading ultimately to the establishment of space settlements," as a national goal, and required periodic reports from NASA on what it was doing to promote those goals, though NASA has dropped the ball on them). And the policy was endorsed again by Presidents Reagan and Bush (the Clinton Administration didn't exactly renounce this goal, but didn't emphasize it, either). But talk is cheap; not much has been done.

What would a space policy aimed at settling humanity throughout the solar system look like? Well, there are some ideas out there, and I'll be devoting some future columns to a more detailed assessment. But the short answer is: Not much like the one we've got, unfortunately.

The most important goal of such a policy has to be to lower costs. Doing things in space is expensive - horribly so. (In fact, in many ways it's more expensive than it was in the 1960s). This is no surprise: it's the tendency of government programs to drive up costs over time, and human spaceflight has so far been an exclusively government run program.

That's why promoting the commercialization of outer space is so important. Market forces lower costs; government bureaucracies drive them up. Among the cost-lowering programs that are likely to make the biggest difference is space tourism, which is beginning to look like a viable industry in the short term. (Just ask Dennis Tito.) We should be promoting it any way we can, but especially through regulatory relief and liability protections.

Government programs should be aimed at research and development that will produce breakthroughs in lowering costs: cheaper, more reliable engines, new technologies like laser launch, etc. Once this technology is produced, it should be released to the private sector as quickly as possible.

Other research should aim at long-term problems: fully-closed life support systems capable of supporting humans for extended periods (you might think that the International Space Station would provide a platform for this kind of research, but it doesn't), exploration of asteroids, the Moon, and Mars with an eye toward discovering resources that are essential for colonization, and so on.

Putting these policies into place would require drastic change at NASA, which is now primarily a trucking-and-hotel company, putting most of its resources into the space station (where the crew spends most of its time doing maintenance) and the space shuttle, which now exists mostly to take people to and from the space station. But we've been stuck in that particular loop for nearly twenty years, without much in the way of results.

It's time for that to change. Like a chick that has grown too big for its egg, we must emerge, or die. I prefer the former.