5 ms·

We tend towards finding smaller and smaller details that we didn't understand as a field matures unless it gets upended by large new discoveries. Classical mech

by rtkwe 3y ago

We tend towards finding smaller and smaller details that we didn't understand as a field matures unless it gets upended by large new discoveries. Classical mechanics for example, we can safely say we've got that pretty much locked down and know what it's good for and where we need to apply other methodologies. Also we have to zoom out from the individual view when thinking about collective knowledge in an entire field. Each individual person has a vast amount they can continue to learn of course but fields as a whole can tend towards more complete knowledge of a particular topic.

tzs 3y ago
> Classical mechanics for example, we can safely say we've got that pretty much locked down and know what it's good for and where we need to apply other methodologies Yet even classical mechanics holds some surprising mysteries, such as how bicycles self-stabilize [1]. [1] https://www.aip.org/history-programs/niels-bohr-library/ex-libris-universum/balancing-bicycles https://www.aip.org/history-programs/niels-bohr-library/ex-l...
- rtkwe 3y ago
  I'm not sure how closely that applies. We didn't, but mostly because people hadn't really studied it and hadn't worked through the mechanics. As soon as we really started looking AFAIK the solution came pretty quickly; it's a combination of 3 things that all make the bike steer towards where it leans: gyroscopic precession, head angle and CoM of the front steering section. All 3 of those things can be tuned to be larger or smaller and that affects the velocity required to achieve stability. https://www.youtube.com/watch?v=oZAc5t2lkvo https://www.youtube.com/watch?v=oZAc5t2lkvo
  - tzs 3y ago
    Note that near the end of that video he says: > But amazingly, even for a riderless bike science currently doesn't what it IS about the special combinations of variables that enables a bike to stay up on its own. We just know that some combinations work and others don't.
    - rtkwe 3y ago
      There have been studies characterizing it more since then both analytically and quantatatively. There have been model bikes built with counter rotating weights to eliminate the gyroscopic effect and it still works. Part of the complexity is they all work together so the parameter space is complex. People have isolated the effect of things like the head angle though or the gyroscopic effect.

by rtkwe 3y ago

tzs 3y ago

> Classical mechanics for example, we can safely say we've got that pretty much locked down and know what it's good for and where we need to apply other methodologies Yet even classical mechanics holds some surprising mysteries, such as how bicycles self-stabilize [1]. [1] https://www.aip.org/history-programs/niels-bohr-library/ex-libris-universum/balancing-bicycles https://www.aip.org/history-programs/niels-bohr-library/ex-l...

rtkwe 3y ago
I'm not sure how closely that applies. We didn't, but mostly because people hadn't really studied it and hadn't worked through the mechanics. As soon as we really started looking AFAIK the solution came pretty quickly; it's a combination of 3 things that all make the bike steer towards where it leans: gyroscopic precession, head angle and CoM of the front steering section. All 3 of those things can be tuned to be larger or smaller and that affects the velocity required to achieve stability. https://www.youtube.com/watch?v=oZAc5t2lkvo https://www.youtube.com/watch?v=oZAc5t2lkvo
- tzs 3y ago
  Note that near the end of that video he says: > But amazingly, even for a riderless bike science currently doesn't what it IS about the special combinations of variables that enables a bike to stay up on its own. We just know that some combinations work and others don't.
  - rtkwe 3y ago
    There have been studies characterizing it more since then both analytically and quantatatively. There have been model bikes built with counter rotating weights to eliminate the gyroscopic effect and it still works. Part of the complexity is they all work together so the parameter space is complex. People have isolated the effect of things like the head angle though or the gyroscopic effect.