4 ms·

Under promise, over deliver… that’s the smart engineering way.

• The_Colonel 3y ago
  No, it's not. As the other commenter nicely explained, this is mostly a byproduct of a very, very low tolerance of failure and extremely difficult conditions, like the impossibility to test it beforehand in the real environment. Under promise, overdeliver, is a nice way to say that you're dishonest with the client. The fact you can overdeliver means you're making some trade-offs elsewhere, e.g. in design/production price, time, weight or somewhere else. Overdelivering is a nice byproduct, but it shouldn't be the goal.
  • lttlrck 3y ago
    When I see it used it's as a device to set expectations when there are so many variables at play that any promise or commitment would be foolish, lead to disappointment with no upside. And that's how I'd interpret it here too. It has nothing to do with dishonesty. Though I have no doubt that it might be used that way, that's fraudulent and a whole different ball-game commercially.
    • The_Colonel 3y ago
      > When I see it used it's as a device to set expectations when there are so many variables at play that any promise or commitment would be foolish, lead to disappointment with no upside. And that's how I'd interpret it here too. In such a situation, why not say like it is? Just explain the client the situation, what's the risk and for what additional cost it can be lowered. And I have no doubt this is what was happening in NASA, in other words this was no "underpromise and overdeliver", this was "explain the situation and agree on trade-offs leading to the best ratio of positive outcome for the price". I think the only situation where this approach is valid is when your client is not sophisticated enough to understand the trade-offs and make an optimal decision.
      • dotnet00 3y ago
        You figured it out yourself, internally NASA is definitely talking about the risks, tradeoffs and more realistic expectations. But officially the statements are always set as an "underpromise and overdeliver" thing because the clients are not sophisticated enough to understand the trade-offs.
        • wolverine876 3y ago
          Do you work at NASA?
          • dotnet00 3y ago
            Nope, just going off of interviews I've seen over the years and just general realities of engineering, kind of hard to design such long lived systems if you're genuinely just aiming for the publicly stated 90 days or whatever else. Plus, I've been learning that the lab I work at is pretty similar. It isn't outright lying, but technically it's not wrong to promise the most pessimistic version of what you think you can deliver that is still interesting enough to get funding. By clients I mean politicians and the general public, who take even delays or testing-to-failure as being actual failure.
            • wolverine876 3y ago
              Thanks, that makes sense.
      • wolverine876 3y ago
        > Just explain the client the situation, what's the risk and for what additional cost it can be lowered. And I have no doubt this is what was happening in NASA, in other words this was no "underpromise and overdeliver", this was "explain the situation and agree on trade-offs leading to the best ratio of positive outcome for the price". That doesn't work well when communicating indirectly, through politicians and journalists, to the public at large. It may not work when communicating to elected politicians.
• dctoedt 3y ago
  > Under promise, over deliver… that’s the smart engineering way. That brought to mind the glass-half-full / half-empty (false) dichotomy: The engineer thinks the glass is twice as big as it needs to be (or maybe 30-40% bigger to allow for a safety margin).