20

Some large community of space enthusiasts is working hard on simulating or extrapolating data on the SpaceX Starship. But this data isn't publicly available.

So why don't they release it as an open source project? Or at least support those who want to create their own open source project?

Machavity
  • 7,905
  • 1
  • 30
  • 69
user721108
  • 4,290
  • 17
  • 32
  • 14
    the boring answer is probably ITAR and wanting to control IP. Not an open source expert but suspect there is a valid answer drawing from successful VS unsuccessful open source projects. In fact has there ever been a successful open source HARDWARE project? Contributing to Starship would really require access to hardware to validate your contribution. – GremlinWranger Jan 26 '21 at 00:14
  • 1
    I've added the business and regulatory tags since as the previous comment points out there are business reasons to keeping some technology proprietary, as well as regulatory reasons for keeping it secret, and making a project open source would challenge those in some ways. – uhoh Jan 26 '21 at 00:28
  • 10
    Has SpaceX ever indicated they want to work with random contributors on this? They seem much more likely to want to control the project end to end, as that gives them the full ability to tailor the direction specifically to that of SpaceX's business plan, surely? Even open source software projects have gatekeepers which keep the project on very tight rails, which is why forks happen and competing development occurs - its not a free for all where everyone wins all the time... If this helps, I can flesh it out into an actual answer rather than a bunch of requests for clarification. – Moo Jan 26 '21 at 03:13
  • 3
    I’m voting to close this question because it's off-topic to space exploration. And almost certainly opinion-based. – RonJohn Jan 26 '21 at 18:18
  • @TooTea As far as I can tell looking at the public data (no insider knowledge of the company) SpaceX is doing a "high tech engineering project" the same (trial and error) way as it was generally done back in the 1950s and 1960s. The rest of the world has moved on since then, because the 1950s way was far too slow and far too expensive. But Musk "gets away with it" because he can - the USA doesn't have any other companies to compare with, and he can waste as much of his own money as he likes without answering to anyone. – alephzero Jan 27 '21 at 11:05
  • 2
    I normally avoid deeps edits like this, but I felt that the rambling verbiage obscured what is a pretty good question. – Machavity Jan 27 '21 at 15:53
  • @alephzero: We can compare it to STS, Ares V, or SLS, for example, and see whether the idea that Starship's development is much slower and more expensive holds up. – Jörg W Mittag Jan 28 '21 at 11:31
  • 1
    "Some large community of space enthusiasts..." employees. They're called employees, and they're paid well. – RonJohn Jan 29 '21 at 05:25
  • 1
    @iono believe what you want, but questions about how VTC works, and where "I’m voting to close this question because ..." comes from belong on meta. – RonJohn Feb 01 '21 at 14:22

4 Answers4

57

There are many reasons. They include money, intellectual property, regulations, and Elon Musk.

Money. Building open source software is relatively inexpensive, sometimes ridiculously inexpensive. No equipment is needed as most programmers have their own computers. There are many open source software projects where the developers do their work for free. Occasionally commercial companies will release some of the software developed in-house as open source, sometimes as a benevolent charitable pay back for all of the open source software they use freely, sometimes in the hopes that external developers will enhance the product, for free. As an example in the space exploration industry, Ball Aerospace made their COSMOS software open source.

Building rockets on the other hand is ridiculously expensive. This is an endeavor that requires very expensive equipment, very expensive facilities, very expensive parts, and rather expensive employees. Elon Musk started SpaceX with \$100 million of his own money. When he showed that he might know what he was doing, he received a downpayment of what would eventually be almost \$400 million from NASA.

That huge investment from NASA was made without NASA getting a stake in the company. NASA was looking for alternatives to the broken cost-plus contracting mechanisms that NASA traditionally used to build new rockets and new spacecraft. That investment was however made with the requirement that the companies that participated in the Commercial Orbital Transportation Services (COTS) project had to obtain significant amounts (hundreds of millions of dollars) of private financing to show the companies had skin in the game. SpaceX did so. Rocketplane Kistler did not, and their COTS contract was cancelled; that company eventually collapsed.

Elon Musk did not use a hundred million dollars of his own money to start an open source project, nor did those private investors.

Intellectual property. One of the key assets of any technology company is its intellectual property, and SpaceX has a lot of intellectual property (IP). Those private investors would have invested zero dollars in SpaceX had Elon Musk said that all of the SpaceX IP would be placed in the public domain.

Regulations. Even if Elon Musk did want to put all of SpaceX's IP in the public domain, there are multiple regulations that would have prevented him from doing so. The US government views rockets capable of going into orbit as advanced weaponry. The International Traffic in Arms Regulations (ITAR) and Export Administration Regulations (EAR) place severe penalties on companies and individuals who release restricted information to other countries, including countries that are US allies. The penalties on companies involve massive fines. The penalties on individuals involve large fines and substandard government-funded room and board (i.e., prison.) Hosting the plans, schematics, flight software, etc. for a rocket on GitHub would result in people going to jail.

Elon Musk. Elon Musk is a strong personality. He does not believe in patents, which is a way of open-sourcing an idea (with a fee). He instead wants SpaceX and Tesla to guard their IP as trade secrets. Musk organizes his companies vertically rather than horizontally as this gives those companies much better control over the supply chain and much better control of its IP. There is no way Elon Musk would have thought for even a second regarding making SpaceX an open source project.

David Hammen
  • 74,662
  • 5
  • 185
  • 283
  • 27
    There is very little difference between the design of a space cargo rocket and an ICBM (inter-continental ballistic missile), which is why the public knowledge of how to build rockets is restricted by international treaty – CSM Jan 26 '21 at 10:38
  • 9
    @CSM Citation needed. There are many international students who study aerospace engineering and related topics in US universities. There are many international members of the American Institute of Aeronautics and Astronautics who receive journals such as the Journal of Spacecraft and Rocketry. – David Hammen Jan 26 '21 at 12:34
  • 1
    Citation to November 2017 handbook: https://www.faa.gov/about/office_org/headquarters_offices/ast/media/export_controls_guidebook_for_commercial_space_industry_doc_faa_nov_508.pdf – Codes with Hammer Jan 26 '21 at 14:13
  • 6
    @CodeswithHammer There is no mention of a treaty in that handbook. The ITAR and EAR are, as far as I can tell, regulations made unilaterally by the US rather than because of a treaty. – David Hammen Jan 26 '21 at 14:18
  • I may have misunderstood what was being queried for a citation in CSM's comment. The handbook I posted addresses rockets being on the US munitions list, independent of the top-level requirement to place rockets on the list. I now see the request for cite was over that top-level requirement. – Codes with Hammer Jan 26 '21 at 14:22
  • 2
    @CodeswithHammer I am not saying that rocketry is subject to the ITAR and EAR. To the contrary! I explicitly wrote in my answer that "Hosting the plans, schematics, flight software, etc. for a rocket on GitHub would result in people going to jail." – David Hammen Jan 26 '21 at 14:47
  • 1
    I agree with this answer, but...open-source anything is not "inexpensive". It costs somebody the time, effort, materials, and other resources used for development. It can be considered "inexpensive" only to those who do not have to bear all the costs, or to those who benefit from the goods or service produced from that project without being directly involved. – Suncat2000 Jan 27 '21 at 13:45
  • @Suncat2000 The vast majority of open source projects (e.g., the vast majority of projects on GitHub) are nearly cost-free. The developers of the typical open source project work for free, at their own cost and using their personal equipment, typically as hobbyists or students, There is of course a cost to GitHub to support all of those public projects; GitHub pays for this by what is essentially a tax on its for-profit private projects. The initial startup investment to start GitHub (Logical Awesome at the time) in 2008 was nowhere near $100 million. – David Hammen Jan 28 '21 at 11:32
  • 3
    @DavidHammen ITAR specifically doesn't restrict sharing textbook knowledge or concepts taught in school but it does restrict sharing of a detailed design. Going from textbook concepts and journal articles to detailed schematics/models for an actual design that you could send out to machine shops and circuit card manufacturers would probably take 10s of thousands of engineering hours. – user4574 Jan 28 '21 at 15:22
  • 1
    @user4574 ITAR cannot restrict sharing textbook knowledge or concepts taught in school, although if the DoD had its druthers, it would mark F=ma as TS/NOFORN (top secret, no foreign eyes). – David Hammen Jan 28 '21 at 21:33
  • 1
    @user4574 When SpaceX started in 2002 it had 150 employees. Assuming 2000 hours per year per employee, that first year alone resulted in 300000 engineering hours. It currently has ~5000 employees. That burns through 30000 to 50000 engineering hours per day. – David Hammen Jan 28 '21 at 21:38
5

In addition to David's answer, I would like to add:

Safety: Developing software for safety relevant use-cases like spaceflights, airplanes, self-driving cars... requires a completely different development approach compared to most open-source projects.

This include the software itself, which must take many regulations into account. They require things like redundant computing, regular system self-tests and so on, which must be planned from the beginning in the software architecture. An open mind community which brings a constant stream of new ideas into the project is not something you want to have here.

Also, the software will require a development process with lots of documentation, testing and certifications to make sure the system will work properly, especially if humans are using it.

Pitto
  • 1,236
  • 5
  • 25
  • 12
    "Open Source" does not mean that everyone can contribute. Some open source projects operate that way, but many others do not. For more information I recommend The Cathedral and the Bazaar. (the author who coined these terms is very pro-Bazaar, but many other projects still prefer the Cathedral model). – Philipp Jan 26 '21 at 10:55
  • 9
    Safety-relevant software is still built from open-source components and that's not a slightest obstacle for having it certified. Someone has to do the validation and testing, but the source can still be public. And some control over what gets added is useful to keep the testing simpler, open-source project can have that. – Jan Hudec Jan 26 '21 at 12:33
  • 20
    This answer is completely incorrect. Nothing about publicising your research inherently makes your project less safe - in fact, for something like this, it could be much to the contrary. Experiments and data can be independently replicated/verified to find more potential errors. Open-source maintainers don't need to accept contributions at a lower standard than the project needs in order for it to be meaningfully "open source". You're conflating "open source" with "amateur" and "illegal". – iono Jan 26 '21 at 15:22
  • @iono Comment only re ". Nothing about publicising your research inherently makes your project less safe " -> Knowing how something works can assist exploits and attacks. While security with total openness is possible, secrecy can help. I suggest that the 2019? Stack Overflow security breach is a salutary example of this. – Russell McMahon Jan 27 '21 at 11:05
  • 5
    Sure, closed source software development avoids all these problems. Two excellent examples was Boeing's MCAS system for the 737 MAX, and before that the "fire-proof" battery management system for the Boeing 787. <\Irony> – alephzero Jan 27 '21 at 11:09
  • 2
    @RussellMcMahon isn't the Stack Overflow security breach actually an example of failed security for closed source software? – Kirk Woll Jan 27 '21 at 17:32
  • @KirkWoll A bit of both (IMHO). The attacker spend days overall reading the blogs and FAQs that explained how the system worked in order to know enough to attack it. They also utilised misuse of the system by the owners in order to exploit weaknesses. A fully open version would almost certainly also have had identical and/or similar mis-uses (read the intriguing blog post for details) and then have been trivially easy to attack. || Where there are people in the loop it's never safe :-) :-(. – Russell McMahon Jan 28 '21 at 00:57
  • 2
    NASA requires that SpaceX disclose its designs, source code, test plans, test results, etc. to NASA for review, and NASA also requires that SpaceX must hire an independent verification and validation (IV&V) organization which also will provided with SpaceX designs, source code, test plans, test results, etc. for review. NASA requires the same thing with other Commercial Crew and Cargo Program Office (C3PO) contracts. – David Hammen Jan 28 '21 at 12:06
  • 1
    Related to safety is the concept of liability. Suppose a rocket undergoes rapid unscheduled disassembly (RUD) and destroys the launchpad (this has happened), or worse, a rocket veers off course and crashes into Bermuda (this has not happened, but an early rocket launched from White Sands did veer off course toward Mexico and crashed a few hundred meters from a Mexican mining company's storage site for explosives). NASA is going to want that launchpad fixed. Loss of life results in lawsuits. International incidents have severe ramifications. – David Hammen Jan 29 '21 at 11:25
  • The FAA requires insurance. SpaceX raised its insurance coverage on a recent Starhopper test to $100 million. The insurance covers the costs of a very bad day. But what if it's a terrible, horrible, no good, very bad day? The FAA wants to see a launch provider with very deep pockets. – David Hammen Jan 29 '21 at 11:43
3

Frankly, until 3d printing gets significantly better (as in print circuit boards and complex electronics better than printers print thermoplastics today), "open source hardware" doesn't really make sense because it doesn't have the key advantages that open source software does:

  • Accessibility: Anyone with an internet-capable device can download an IDE, git, and a programming language and go to town. This means that a 16 y/o schoolgirl living in New Delhi can contribute just as much as a centenarian retiree who lives on a ranch in Indiana. In contrast, even the simplest open source hardware projects require manufacturing capability which very few people have. To get started with even the low end of projects, you're going to need to shell out at least several hundred on a 3d printer.
  • Individual skill: Writing code is both easier to learn and easier in general than engineering a physical product. Anyone can access a free online learn-to-code course and with zero investment besides time, become a proficient programmer. Engineering on the other hand basically requires higher education or some sort of apprenticeship program.
  • Iteration speed: The iterative cycle on writing code is potentially in the seconds range. I can type some code, click compile, test it, and be back to writing code in several heartbeats. Engineering on the other hand requires a longer iteration cycle. Even in "rapid prototyping", you're still waiting hours minimum for a 3d-printer to spit out a part and if you include more complex manufacturing processes (electronics) you're waiting days for shipments and spending hours soldering.
  • Collaboration ease: With version control software like git, multiple coders can easily work on the same project and then merge their results afterwards to combine their efforts. This is significantly harder in engineering as you can't "merge" drawings, sketches, or CAD files (collaboratively making an assembly works but is difficult). Additionally, to properly work on any hardware project, you need the physical prototype to test and experiment with. This means that anytime anyone in the project team makes a change to the master branch, everyone would need to re-download and re-manufacture the latest changes.
  • Professionals and money: Frankly, anyone good enough at engineering to contribute to a cutting edge project like Starship is either a) Already employed or b) Unemployed and looking to get paid for work.
  • Individual benefit: Open source works because individuals want to use it. OpenHab (a home automation software) is used by people who want to automate their home, and when they find software glitches in the system, the fix and share them because this benefits them directly. Working (for free) on a project like Starship brings no at-home or future benefit to the contributor (outside of maybe feeling good about it).
Dragongeek
  • 18,658
  • 2
  • 63
  • 109
  • The people working for Relativity Space would strongly disagree with your conjecture regarding 3D printing. They think that 3D printing is a very good idea for making rockets. Of course, their 3D printing technology is not open source. It is their central piece of intellectual property, and they presumably guard that IP very closely. – David Hammen Jan 30 '21 at 09:45
  • @DavidHammen You're right, but my point about 3d printing is not about the actual additive manufacturing process but rather the democratization of at-home making that consumer grade printers have born over the last decade. Even if Relativity Space's printer can print a rocket, people don't have a Relativity Space-quality printer at home so they wouldn't be able to contribute to any open source design. – Dragongeek Feb 02 '21 at 14:09
  • Downvoted because this seems to focus too much on spurious arguments against open source hardware (which is a thing, does work, and I've taken advantage of it)... and not enough actually talking about what opensource would mean for starship. – NPSF3000 Feb 02 '21 at 16:43
-4

The answer is very simple: capitalism.

The implicit and/or systemically-rewarded goal of any privately-owned (as opposed to democratically-controlled, not state-owned or publically-traded) company is to profit as much as possible. SpaceX publicising the data from its development efforts could potentially benefit other companies or government organisations that might compete with SpaceX; SpaceX is using "closed-source" / intellectual property laws to retain its "first-mover advantage".

iono
  • 109
  • 3
  • 1
    That doesn't explain why a large number of open source projects have been developed and thrived within a capitalist society of the USA and Europe but this one hasn't been open source. – ikrase Jan 28 '21 at 04:11
  • @ikrase Open source can be profitable to certain capitalists in certain circumstances, but not others. It was capitalists who actually supported legally abolishing child labour in Britain during the Industrial Revolution: an individual capitalist would lose out in the short term by abstaining from child labour, but if they collectively did it through the state, it was in their collective class interest: preventing workers from en masse early deaths made them, on the whole and in the long term, more productive. Sometimes "the right thing" aligns with profit, sometimes it doesn't. – iono Jan 28 '21 at 07:56
  • 2
    "In SpaceX's case, it is a publically-traded company" – Citation needed, as they would say on Wikipedia. – Jörg W Mittag Jan 28 '21 at 11:18
  • 2
    "In SpaceX's case, it is a publically-traded company" -- This is not the case, so very much not the case. Elon Musk has intentionally made SpaceX not be a publicly traded company. It is not listed on any stock exchange. It is instead a privately financed company. Elon Musk does not want to even chance being overruled by a SpaceX board of directors or by the various public meetings a publicly held company must endure, and he especially does want to chance being kicked out as the CEO. This answer is flat-out wrong. – David Hammen Jan 28 '21 at 11:40
  • I'd upvote this if you corrected the "publicly traded company" thing. – WaterMolecule Jan 28 '21 at 19:16
  • @DavidHammen you don't seem to have understood the rest of the answer, given your "flat-out wrong". Privately financed companies are a subset of privately owned companies, so all the properties of privately owned companies relevant to answering this question apply to SpaceX. The difference between "a single autocratic self-interested decision-maker" vs "a group of autocratic self-interested decision-makers with fluid membership" is totally immaterial in the context of systemic behavioural pressures. – iono Jan 31 '21 at 10:43
  • @iono I had not yet downvoted when I made that comment (but now I have). My downvote is for the initial sentence, The answer is very simple: capitalism. There is no doubt that Elon Musk likes being very, very rich. (continued) – David Hammen Jan 31 '21 at 11:17
  • However, there also is no doubt that Elon Musk wants to send people to colonize Mars, including himself. If doing so makes him poor (poor in the sense of no more fantastically cool parties), so be it. He sees himself as saving humanity by spreading it beyond planet Earth. In his mind (he has said and written this), he must be one of the colonizers to insure that this will be successful. – David Hammen Jan 31 '21 at 11:17
  • @DavidHammen You still don't seem to understand the very basic subject of my answer. I'm talking about system behavioural pressures. You're using the public statements of a single billionaire, who has a substantial vested interest in appearing benevolent in motivation, to conclude that they are indeed benevolent in motivation. The Starship is closed-source because the Starship is entirely controlled by a single capitalist. The Starship is entirelly controlled by a single capitalist because capitalism produced these circumstances. – iono Feb 01 '21 at 07:21