Looking to the Future -- The Limits of "Move Fast and Break Things"
First, I'm not knocking SpaceX here. They did a lot of hard work, and learned a lot of things.
But looking to the future -- the far future of FTL travel and space colonies -- we're going to have to find and work in a mode of ultra-conservative, get-it-right-the-first-time spacecraft/spacestation development, 180° opposed to the current general-software and automobile-software development mode of "We'll just patch it and ship an OTA update." The reasons being:
* Interstellar spacecraft will be huge capital investments, FTL craft even more so. Time is money, and the craft owners won't want to lose money while their expensive ships are tied up at dockside, getting yet more refits and upgrades. The owners will want those ships serviced and back out in space, carrying goods, passengers, and information.
* You don't want docking failures because some local-space shithead politicked and got computer interface protocol, or docking collar design (which his large contributor builds) used in his local-space vicinity "upgraded" from spec V1.1 to V2.3. The ship(s) can't just be upgraded (even if they wanted to) to V2.3, because then they could no longer dock with stations running V1.1. Yes, you can EVA your passengers from ship to station, but EVAing cargo modules will be slow and costly, even were it possible. More money lost there. Eventually the word would get out among the ship crews, and those ships will no longer stop at Space Station Y, which would then feel economic damage from the lack of ship traffic.
* Backward-compatibility and adaptors might-possibly be used, but they add complexity and reduce reliability.
* Interstellar space comms are slow and unreliable; they're not what a wise and moral person would want to use to update human-safety-critical systems.