‼NOTES‼ 1. First half of the video is decent. Second... I was already 'brain dead', but too stubborn to stop. Probably not worth to watch... 2. The function scoring the path needs improvement. I 'think' dividing the path based on the two extremes on x is something that is obvious but I didn't try for some reason. Then the score could be the ratio / difference of the top to bottom areas. 3. The 'Leg' should **not** have a motor. But it should provide endpoints to attach to a motor. 4. The 'Strandbeest' needs its own object. It has a motor and multiple legs attached to it. It can use the 'Leg' object to create the sticks and the helper functions to attach to the motor, but then it needs its own update method. Now each leg is updated independently, so, updating the second leg, breaks the alignment of the first one (I think). 5. The physics needs friction with the ground. I would try to see if the particle is touching the ground before and after updating (it and the sticks). If the particle is still touching after all that, the particle is reverted to the previous position, made static and sticks are updated again. More Thoughts... 1. I think polar coordinates wouldn't have been a good idea after all, because we don't know the angles between those points, so the calibration technique still makes sense. 2. I think inverse kinematics wouldn't work, because this skeleton can only reach points on the generated curve. it's essentially a way to map the circle on another very specific curve, so, the tip cannot be forced to any position outside that curve without breaking the segment. ‼CHALLENGE‼ I don't plan to touch this code until March. If someone manages to improve it, I can dedicate a live stream showcasing their solution and doing few more experiments with that. The other reason I couldn't stop coding was that it was so captivating to see it move :-) don't mind seeing more of that!
Radu Mariescu is truly remarkable, inspiring new developers with his brilliant thinking process and guiding them step-by-step on how to tackle even the most complex projects with ease. Thank you 🙏 for you time.
Insane, you coded for 9+ hours straight to make it work, I watched 2 hours in start only and i think i will take your challenge and try atleast but before that i will more fine tune the ball game.
‼NOTES‼
1. First half of the video is decent. Second... I was already 'brain dead', but too stubborn to stop. Probably not worth to watch...
2. The function scoring the path needs improvement. I 'think' dividing the path based on the two extremes on x is something that is obvious but I didn't try for some reason. Then the score could be the ratio / difference of the top to bottom areas.
3. The 'Leg' should **not** have a motor. But it should provide endpoints to attach to a motor.
4. The 'Strandbeest' needs its own object. It has a motor and multiple legs attached to it. It can use the 'Leg' object to create the sticks and the helper functions to attach to the motor, but then it needs its own update method. Now each leg is updated independently, so, updating the second leg, breaks the alignment of the first one (I think).
5. The physics needs friction with the ground. I would try to see if the particle is touching the ground before and after updating (it and the sticks). If the particle is still touching after all that, the particle is reverted to the previous position, made static and sticks are updated again.
More Thoughts...
1. I think polar coordinates wouldn't have been a good idea after all, because we don't know the angles between those points, so the calibration technique still makes sense.
2. I think inverse kinematics wouldn't work, because this skeleton can only reach points on the generated curve. it's essentially a way to map the circle on another very specific curve, so, the tip cannot be forced to any position outside that curve without breaking the segment.
‼CHALLENGE‼
I don't plan to touch this code until March.
If someone manages to improve it, I can dedicate a live stream showcasing their solution and doing few more experiments with that. The other reason I couldn't stop coding was that it was so captivating to see it move :-) don't mind seeing more of that!
Radu Mariescu is truly remarkable, inspiring new developers with his brilliant thinking process and guiding them step-by-step on how to tackle even the most complex projects with ease. Thank you 🙏 for you time.
Thanks, but 'with ease' doesn't really describe what happened yesterday :-))
Insane, you coded for 9+ hours straight to make it work, I watched 2 hours in start only and i think i will take your challenge and try atleast but before that i will more fine tune the ball game.
No point to watch more than 4 hours, I think. I was really tired in the second half.
Good luck 🙂I added some details in the pinned comment.