A Russian Soyuz rocket launching a scientific satellite exploded half a minute after launch on Tuesday.

It carried no crew, but it was of a similar type (though out of a different launch site) to the vehicle that delivers Russian cosmonauts (including paying space tourists, such as Dennis Tito and Mark Shuttleworth) to the International Space Station. Had the finances worked out, teen-heartthrob Lance Bass would have been aboard one this month.

The jury is still out as to whether or his loss would have been a net tragedy. For listenable music, perhaps no, but for his friends and family, and the future of spaceflight, almost certainly…

Of course, the vehicle that was planned to carry crew was built to more exacting specifications, so it’s possible, even likely that this particular failure wouldn’t have occurred.

And had crew been aboard, there’s still a reasonable chance that they would have survived, because the manned Soyuz launches have an escape tower in the capsule, which allows the crew module to be rapidly carried away from a disintegrating launch vehicle. This was demonstrated about nineteen years ago, when the Soyuz T-10A mission saved its cosmonauts in this fashion.

But even if the vehicle is “safe” in the sense that the crew can survive such catastrophes, its reliability would be considered unacceptable for any other mode of transportation. And the cost would still be horrendous, because the mission would be delayed, and another vehicle would have to be prepared and flown to complete it.

While the version that failed wasn’t “man rated,” and this rocket is one of the most highly-used and reliable in the world, the failure points up the continuing folly of relying on converted munitions for delivering people to space, which some NASA plans would seem to perpetuate. Unfortunately, many people still propose, as the future of US manned space transportation, putting a reusable module on top of an expendable launch system.

The failure of expendable launchers isn’t the fault of engineers. Or at least, it’s not a result of bad engineering per se (though it may be a result of bad engineering management judgment). It’s an intrinsic problem with expendable launch vehicles flown at low flight rates.

The argument over reusable versus expendable launch systems generally is an argument of economics. At least for high flight rates, reusable systems, in theory, should be much less costly because the main costs are the consumables (primarily propellants) whereas in the case of the expendable launcher, the entire vehicle has to be replaced each time. But reliability is an issue beyond cost (though in fact, it’s a part of cost, because failures are costly in themselves).

Can expendable systems be made reliable? Yes.

Soft-drink and beer containers are expendable, they are produced and used in the millions, they are light weight, and they rarely fail.

But expendable rockets are not produced in those numbers, and they consequently don’t have a high enough production and flight rate to develop the kind of statistical quality control that, say, bean canneries have. There may be a market for which throwing rockets away can provide both low costs and high reliability, but we haven’t seen it to date.

And if such a market exists, I suspect that it will be better satisfied by a reusable system. Reusable space transports (that is to say, truly reusable rockets, which have never been seen to date–Shuttle doesn’t really count, because much of it is expended, and the rest is extensively rebuilt between flights) have a much higher potential for not just low per-flight costs, but very high reliability.

The reason for this is that a truly reusable vehicle has a flight history. Every flight of an expendable rocket is the first flight. No matter how many have been built, each one is new, and may have manufacturing problems that cannot be wrung out, because each flight of the vehicle is not only its first, but also its last. It suffers from what engineers call the “infant mortality” problem.

Imagine that every time you went to the store, you did it in a different car. Even though the quality at the factory might be high, you’d still have more confidence in a car that you’d driven previously without incident, than one that had never been driven anywhere. Yet that’s what expendable vehicle manufacturers are asking us to buy–a virgin system that has never had its integrated systems tested as one, let along flown.

But getting back to the cost issue, there is no other mode of transportation in which we expect to throw the vehicle away with every trip. If we did, we certainly wouldn’t expect to pay a low cost for the ride–we all know how much transportation hardware can cost. As an example, consider the Boeing 757. The current price for such an aircraft is about eighty million dollars, ignoring various options. It has a capacity of about 200 passengers.

Now imagine that we flew cross country with it, which in fact many do. If, at the end of the flight, the airplane was taken out to the end of the field and blown up, the lowest possible ticket price that would allow an airline to break even (forget about profit, or even the other costs of running an airline of the marketing, kerosene, crew and various overhead functions) would be $80M/200 or $400,00 per one-way ticket. Even ignoring the reliability implications that arise from the fact that every flight is a first flight, does anyone think that there would be anything resembling the present airline industry at prices like that?

Yet in space transportation, we accept this as a normal state of affairs. As long as we continue to do so, space will remain a sterile place, empty of not just human life, but as far as we know, all life. It will continue to be reserved only for the very wealthy or very well connected, and for the most part, irrelevant to the lives of everyday people.