"Regulatory" isn't the right word, but I don't know what is.

For example, SpaceX seemed to be able to prove to NASA's satisfaction that they were sufficiently safe to use for crewed launches in 2020. The very little I know of SpaceX's processes and approaches suggest that they would not have aligned to NASA's systems engineering process very well, so how were they able to convince NASA they were ready for manned flight so quickly?

I am especially looking for a process description here.

  • 2
    $\begingroup$ The CCDev (commercial crew development) process was based on the COTS process that preceded it. There were still years of meetings doing requirements development (I was in the rendezvous ones alongside NASA EG and others), and an extended feedback loop and iteration between SpaceX and NASA. I'll try to figure out a way to put together a proper answer; it seems like at least some of the documents are public, though imo the most important one, SSP 50808, is ITAR controlled :( $\endgroup$
    – Erin Anne
    Commented Mar 16, 2023 at 23:07
  • 6
    $\begingroup$ I wouldn't call it "quickly" as SpaceX was founded 20 years ago. The company has been quite Agile from the start. It took a good amount of time for SpaceX to convince NASA that Agile was a "good thing". NASA on the other hand had expected a waterfall model, even though NASA knows waterfall does not work. $\endgroup$ Commented Mar 17, 2023 at 3:20
  • 4
    $\begingroup$ Many, many good things have been delivered using waterfall @DavidHammen, including Apollo. $\endgroup$
    – GdD
    Commented Mar 17, 2023 at 9:12
  • 1
    $\begingroup$ @ErinAnne I would very much like to read that answer. That spec is a great example: I would expect demonstrating compliance with it to be a significant undertaking. I also would be surprised if mere spec compliance satisfied whatever safe-for-flight process NASA uses. If there is a good point of contact for requesting the 50808 (not familiar with it) and that's the meat of the answer, I can send a request offline that unambiguously complies with ITAR. $\endgroup$
    – fectin
    Commented Mar 17, 2023 at 14:05
  • 5
    $\begingroup$ @fectin There's an issue beyond being qualified to handle ITAR data in a general sense, which is "need to know." The rationale is that there's not a much of a difference between knowing how to launch and land a civilian spacecraft and knowing how to launch and land an intercontinental ballistic missile, and there's not much of a difference between knowing how to slowly rendezvous with the ISS and knowing how to violently rendezvous with (i.e., destroy) an enemy spacecraft. Launch, landing, and rendezvous are about as dual use as technology can get. $\endgroup$ Commented Mar 17, 2023 at 20:34

1 Answer 1


First, it is because the Falcon 9 have a good track record. Much of regulation is basically convincing a whole lot of people that we know what we're doing and we're not going to cause a problem for you. By 2020, SpaceX had been fighting through what they perceived to be mostly useless bureaucratic regulations. SpaceX had a pretty good track record of 67 straight success with the Falcon 9 and had done the CRS and COTS program. During those programs they had only one failure, back in 2015, so they started off with a higher level of trust. By contrast, take Starship. Look at how much regulation issues that is facing. I won't go into it here, but you can search up "Why hasn't start gotten an FAA launch license?".

Additionally, one test flight is not exactly what happened here. While Dragon is different, the CRS and COTS program missions were all test flights of a sort. For the launcher, all 84 (85 counting the AMOS-6 on ground failure) missions were tests. The only thing Crew Demo-1 was testing in one flight was the things required for humans.

Finally, one test flight isn't all that crazy either for a reliable launcher:

  1. Starliner only planned to have one test flight (then the test flight failed).
  2. The SLS only had one test flight before it will have crew.
  3. The shuttle had 0 uncrewed test flights (I won't go into why).
  4. The Titan II GTV (the Gemini vehicle) only had 1 test flight to verify it's safety for crew (the other was solely to test a heat shield)

I could continue, but you get the point. This isn't unheard of.

  • 2
    $\begingroup$ Unfortunately, I think this is a great answer for a different question. I really do want to understand SpaceX's process here, and am far less interested in whether that process was appropriate or safe. $\endgroup$
    – fectin
    Commented Mar 17, 2023 at 14:04
  • 4
    $\begingroup$ @fectin I agree with Starship is go for launch: Their process gives them a competitive advantage and hence SpaceX treats it as a trade secret. SpaceX's IV&V contractor knows their processes, as does NASA, but since it is a trade secret, you will not find it. I work for SpaceX's IV&V contractor, which also is the same contractor NASA uses for certifying that SpaceX has done the right thing. We are firewalled to the max. As I know other companies trade secrets, I cannot work with either of those two groups. That means I am free to say what is in the media, which is that SpaceX is very Agile. $\endgroup$ Commented Mar 17, 2023 at 15:47
  • 4
    $\begingroup$ This answer needs sources badly. It's also incorrect about "only one failure" for Falcon 9; CRS-7 exploded during ascent. That was the 2015 failure; AMOS-6 was in 2016. $\endgroup$
    – Erin Anne
    Commented Mar 17, 2023 at 18:44
  • 4
    $\begingroup$ While not part of CRS, AMOS-6 occurred between CRS-9 and CRS-10, and 100% affected the amount of trust NASA had in SpaceX. $\endgroup$
    – Erin Anne
    Commented Mar 17, 2023 at 18:54
  • 2
    $\begingroup$ The Shuttle did have a crewed test flight. The reason it had no uncrewed test flights was that the Shuttle needed a human pilot to land. The Shuttle did not have an automated landing capability. $\endgroup$ Commented Mar 17, 2023 at 20:09

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.