The first statement implies that computer testing is so good now that an aircraft can be tested so thoroughly that the maiden voyage of an aircraft could safely also be its first "for real" mission.

It is true that computer models have definitely reduced the risks for test pilots by helping to define a much greater percentage of the performance envelope of a given system without the need for test flights. When a pilot steps into an untested craft for the first time, the "unknown" area of the envelope is better quantified; in different words, we have a better idea of how much of the performance envelope is unclear.

Yet that unknown area still exists and has to be discovered by actual human test pilots because as good as computerized testing is there still isn't 100% coverage.

(I'm a pilot and have dealt with experimental aircraft. Not on the level of this project, of course, where the envelopes of technology are being pushed, but on the level of small private aircraft where the technology is in many cases half a century old, well understood, debugged, and still has to be tested. By a human.)

I think he meant real-life test, but with the pilot working remotely and not sitting on board.

there's a difference between testing the engine, and testing the handling qualities. But not always.

For testing the engine which in this case is throttleable and has time constants greater than hundreds of milliseconds the time lag from not having a pilot onboard is acceptable.

For flying qualities again the plane is large enough to be slow enough that remote pilot timelag is not significant.

There is no part of the envelope of this vehicle where you need a pilot onboard. Most of the development here can be done without pilot onboard. Space has been done from unmanned vehicles since the beginning and what they are doing is only barely rocket science.

Putting a pair of pilots in a vehicle with a non-flight-tested rocket engine is just irresponsible.

I like Rutan's company. But what they are doing on this program smacks of amateurism. They got engineers killed on the ground and now pilots killed and injured in flight. Both of which were completely preventable by some basic range and flight safety measures.

It looks like they are cutting corners by doing too much too soon because of management pressures to cut delays. And people are paying the ultimate price for their hubris.

Your forgetting the countless times that a test pilot has saved the aircraft/spacecraft from destruction. If it was unmanned and flown by computer, it would have been a fireball.

At least one shuttle flight would have been lost if the crew didn't fly manually to correct for a malfunction. STS-4 PIO'd on landing and would have crashed if auto-stabilization wasn't overridden by the pilot.

According to wikipedia, it was STS-3 that had (autoopilot-induced) oscillations during landing. The description on Wikipedia seems to imply that they were expected (the autopilot was being tested) and that the dangerous condition was due to the autopilot being left partially engaged after the test.

Correct. There was also an pitch stabilization system that malfunctioned due to the misconfigured (partially engaged) autopilot. Point remains that the pilots flying skills in the cockpit saved the orbiter.

And in this case the craft killed the pilot. Shouldn't we be designing pilotless craft?

The last of the ALT drops did this (with Fred Haise as PIC, IIRC), as did STS-3 back in 1982 at White Sands. I'm not familiar with the STS-4 incident at all. Do you have any links?

Pilots can make manual correction remotely. It might not be optimal, but that is better than risking lives on the first flight of a new engine design.

It was STS-3 at White Sands, not STS-4, my mistake.

not in this case. They should have more than one rocket craft for testing. it's ok to fly reciprocating and jet prototypes because Scaled has experience with hundreds if not thousands of those kinds of aircraft, and it's easy to bail out at 200mph. It's much harder to bail out at Mach 3 and 100k ft.

>there's no need not in this day and age to have test pilots.

Good point. What does the FAA say about unmanned rocket powered flight tests?

If a regulation is putting human lives at risks, the regulation has to go.

Not sure, but SpaceX has performed plenty of them.

Not so much over the US, though.

I would beg to differ. SpaceX certainly flown unmanned, rocket powered test flights both from McGregor[1], and Cape Canaveral[2]. Those were in the US last time I checked.

1:http://en.wikipedia.org/wiki/Grasshopper_(rocket)#Flight_tes... 2: http://en.wikipedia.org/wiki/Falcon_9#Launch_history

plenty of companies do them. Ask Elon Musk.

Did you know an F1 car can drive itself around the track? It seems like there's something in human nature, at least among some people that leads them to want to drive the thing themselves.

not today's F1 car. They don't have auto trans, auto steering or auto brakes. Yes they could and audi has been testing an unmanned race car that does round a track, but there is a little chance that an F1 engine will blow up and kill the driver.