> Here's why it's likely the manned Dragon capsule will be one of the safest manned spacecraft in history.

Lets not forget that the two shuttle accidents were as much about a human factors failure, as a systems failure. Make no mistake about it, If a company like SpaceX ever got complacent about safety, it is game over. They will not get a second chance. Remember McDonald-Douglas, and how many DC-10 accidents, that basically forced the company out of the airliner business.

The Launch Escape System on a Soyuz capsule has been used once, in 1983 http://en.wikipedia.org/wiki/Soyuz_T-10-1

The Space Shuttle was a far less reliable system from a fundamental design perspective. The size of the wings, the nature of the thermal protection system, the insulating foam and the use of a super-cryogen as fuel, the side-by-side staging, the use of solid fueled boosters on a manned launcher, etc. Engineering estimates show that we were incredibly, incredibly lucky to have the Shuttle safety record we did, the vehicle was a death-trap. We almost lost Atlantis in 1988 and Discovery in 2005 to the same thing we lost Columbia to, and we almost lost several shuttles in the early 80s to the same fate that doomed Challenger.

Of course if an organization ever gets complacent about safety then things can go downhill, but in the case of SpaceX they have a fundamentally more robust vehicle design, a system with more and safer abort scenarios, and every possible reason to maintain a high degree of vigilance about safety (if for no other reason than their company's image and financial bottom line).

Saying "we almost lost Atlantis" only confirms the point that they could have mitigated the foam lost before Columbia, and they could have fixed the O-Rings before Challenger. Not investigating those incidents properly, is a human factors failure.

I do agree that design decisions like side-by-side staging of SRBs wasn't the safest choice. Safer abort scenarios is definitely a feature of Dragon.

One problem i see with spacex is their propulsive landing feature. What if things go wrong at that point, like a software or a sensor problem causing it to descend at a speed too high/not realizing it approaches the surface/just hovering and expending the propelland quickly? If landing on land, they don't have any backup, and astronauts are doomed, falling from even very low altitude (30-50ft) will be fatal.

Remember the space shuttle is a 1970s design. I'd certainly expect that a 21st century design would fare better in nearly every regard.

To some degree that's the case. The Falcon 9/Dragon do very much get a benefit from modern technology. Better machine tools, high-speed digital communications and onboard computers, better material science, the ability to run computer simulations of different flight profiles or to model the operation of a rocket engine, etc. However, in another sense the Falcon 9/Dragon represent very primitive designs. Indeed, 1950s or 1960s designs even. A simple cylindrical 2 stage LOX/Kerosene booster in combination with a simple frustrum shaped manned capsule. These designs have heritage going all the way back to project Mercury.

Yeah, and guns are vastly more reliable than rockets or missiles, too. Sometimes old and simple is the way to have high reliability, sacrificing performance.

The space shuttle was a poor design in many respects even by the standards of the 1970s as described in Feynman's Challenger crash report http://www.ralentz.com/old/space/feynman-report.html

Good design does magically appear as time goes by but is a result of a sound design and testing process. Fortunately, I am sure SpaceX is aware of this so they will most likely get a much better track record than the space shuttle.

why?

Because we have new materials, new technologies, and 40 years of experience with space plane design and operations, that's why.

The U-2 spy plane is in service after the retirement of its original replacement.

The B-52 is still a heavily used part of the US heavy bomber fleet and current plans are to keep it in service until 2040 and perhaps beyond (which is longer than the B-1b is planned to be in service).

I wouldn't say the DC-10 problems of the 1970s forced McDonnell-Douglas out of the airliner business. They continued in business into the 1990s before being acquired by Boeing. The DC-10 itself was in passenger service until just a few years ago, and is still flown in freighter and military configurations.

The DC-10's legacy is really one of bad timing more than anything. Its direct competitor, the L-1011 suffered a very short production run for many of the same reasons, none of which involved safety.

The 1980s ushered in a new era of twin engined widebody aircraft and the phasing out of the flight engineer. These two trijets were consequently a difficult sell, although the MD-11 was competitive with the 747 briefly in the early to mid 90s.

The last DC-10 airliner was made in the 80s. The 747 is still in production. People aren't buying tickets to fly FedEx so the safety perception isn't as big an issue.

Perception is the key, statistically the DC-10 was not an unsafe aircraft. Its hull loss and fatality rates are not much different from the 747.

Statistically the 747 is 6x safer than the DC10, and its figures are distorted by the Canary islands collision that had nothing to do with airworthiness.

http://www.fearofflying.com/resources/safest-airliners-and-a...

That said, today both planes are (literally) perfectly safe.

The DC-10s problems all stemmed from a bad cargo door latch design combined with insufficient redundancy in its hydraulics.

> More so the vehicle is designed to be able to withstand an engine exploding, even at launch, without damaging the other engines and without preventing the launcher from reaching orbit.

That is absolutely incredible. Gives new meaning to the phrase "over-engineered" (and not in a bad way).

Do you know off the top of your head what's the failure rate for a generic launch? How much is it saying that the Falcon is at 100% so far?