Starship: The Iterative Design Methodology

22 April 2023

Starship: The Iterative Design Methodology

By Gwynne Dyer

“Obviously, this is not a nominal situation,” said John Insprucker, a senior engineer at Space-X, who was doing a webcast on Thursday’s launch attempt of Elon Musk’s gigantic Starship rocket. So why did Musk’s employees, hundreds of whom were watching live, cheer when it blew up only four minutes into flight?

Because it made it off the pad. In fact, it reached 39 km. high before the range safety system blew it up: not bad with at least six of its 33 engines out of commission. It was never intended to land again. The plan, if it got that far, was for both components, the Super Heavy first stage and ‘Starship’, the payload, to crash into the sea.

This is actually how Musk’s development strategy, also known as the ‘iterative design methodology’, is supposed to work. You push the system under development until something breaks; you figure out what went wrong and fix it; then you push the system further until something else breaks. Repeat until nothing breaks, and you have a system that works.

Large, government-funded organisations like NASA can’t do this. They have to sit around for years exploring and hedging against every imaginable risk, because failure is a big black mark against them. Whereas in the iterative model, failures are how you advance.

The iterative strategy may cost more, at least in the short term, but Musk can afford it. You can’t do it with people aboard – but ideally you won’t put crew aboard until you have found and fixed all the failures. And it is fast: Musk first mentioned the idea for Starship only ten years ago. (That’s very fast, as these things go.)

So all the people who actually have money on the table (including NASA, which is depending on Musk to provide the lunar lander) are quite happy with his progress. All the more so if it turns out, as seems very likely, that Thursday’s problem was not with the rocket at all, but with the launch pad.

This is the biggest rocket ever launched: twice as powerful as its nearest rival. The blast from its 33 engines was enormous, and for fully 8 seconds it was clamped to the pad while its engines wound up to full power. In the videos, you can see the concrete of the pad disintegrating and flying around, and a hole being carved into the earth beneath it.

They’re working on a water deluge system to protect the tower, but it isn’t ready yet. Usually there’s a flame trench to divert the rocket blast away from the pad, but they thought they didn’t need it. The debris flew up and struck some engines on one side of the cluster, which flamed out. That unbalanced the thrust, and the rocket began to tumble. Boom!

Space-X’s engineers are not infallible, but they’ll just try again. Three more Starships are already built, and Musk says they’ll test-fly another one “in a couple of months.” If the damage to the pad is bad enough it may take longer than that, but it will still probably be this year.

And what makes the partially successful test flight of a new rocket in south Texas worthy of more than 500 words in this estimable publication? The fact that this rocket is how the human race gets out into the universe. Maybe the only way.

Starship will cut the cost of getting material into low Earth orbit a hundredfold. It can refuel in space, and carry really big payloads to most parts of the Solar system. It is the game-changer: once this technology is fully operational, we are out for good. And what Elon Musk is doing now could have been done forty years ago.

There’s no ground-breaking technology in Starship that wasn’t there or almost there by 1983, except some aspects of computing power. People just got distracted by wars and other crises in the 1970s, and the whole enterprise of space got shunted onto a siding. That could happen again.

The enthusiasm is back now, but even in this era Musk is the only person with the drive, the imagination and the resources to build something like Starship. Neither the man nor the era will last forever, and it could all slip away again if the foundations of a space-based economy (which requires Starship) have not been laid by the time one or the other is gone.

Musk suffers from random enthusiasms (like Twitter) that leave him exposed to financial disaster. Some foolish war could draw in the great powers and turn everybody’s attention elsewhere. Some sudden lurch in the climate could leave us all struggling.

But if none of those things go wrong, we will be a genuinely space-faring species in ten years. Fifteen, at the outside.


To shorten to 700 words, omit paragraphs 4 and 5. (“Large…go”)

Gwynne Dyer’s new book is ‘The Shortest History of War’.