4 ms·

Im not sure that RD works that way. Innovate, time and compatibility, pick two.

• DanielBMarkham 8y ago
  I don't know. I know that some mention was made of the pieces fitting already-existing army wheel specs. So obviously for some groups integration was a consideration. And it's a mistake to think in all-or-nothing thinking here. I didn't say you needed to design a rock-solid integration system, only that it looked like integration was a minimal concern -- and that was an interesting choice. I come from the tech community. The longer we put off integration, the more expensive things become.
  • dannypgh 8y ago
    Putting off integration is cheaper when it's not yet clear what you need to integrate with. Or to frame it differently: having requirements of "be everything for everyone" is a surefire way to fail.
  • hedvig 8y ago
    This sub-thread reminds me of the monolith vs microservices kernel discussion I see a lot here. Doesn't Linux benefit from specialized pieces being piped together at will?
    • DanielBMarkham 8y ago
      Yep. It's a spectrum and a scale. On various scales you want to apply various paradigms. It's not a yes/no thing. We know for a fact that no matter what they pentagon does with wheels, they need to fit on a body and cost less than a million bucks or so. Even giving the R&D teams wide latitude for they're playing around in, we can start setting some reasonable guardrails up.