3 ms·

> An individual is never going to invent A* on the spot without an education, for example. What? A* is a brute force solution. Here's a complete description of

> An individual is never going to invent A* on the spot without an education, for example. What? A* is a brute force solution. Here's a complete description of A*: A* is a breadth-first search in which we use weighted path length, rather than unweighted path length, to calculate tree depth. Tons of people will invent it on the spot. The algorithm is just "let's try everything, in order of desirability, until something works".

zarathustreal 2y ago
Look brother you’re not winning any points by implying that you’d be the guy to invent A* on the spot. Without formal education you’re almost certainly not even dealing with concepts like breadth-first search, trees, or nodes, so no, a ton of people will not be inventing it on the spot. Edit: most people don’t even think about recursion as a possibility until taught about it in a formal setting, and even then it’s typically a struggle to understand. Even using a queue, that’s a data structure that a person with a task to do isn’t likely to spend the time to invent on the spot.
- hnfong 2y ago
  I agree the vast majority of people aren't going to invent A* on the spot. But I don't understand your emphasis on "formal" education. I'm pretty sure a sufficiently smart high school kid interested in the subject could learn all these topics online in a couple months without enrolling in any college.
- thaumasiotes 2y ago
  > Without formal education you’re almost certainly not even dealing with concepts like breadth-first search, trees, or nodes, so no, a ton of people will not be inventing it on the spot. If you're not dealing with nodes, you won't be inventing A* on the spot because you don't have a problem to which it can be applied. But otherwise, it's literally a breadth-first search with the queue replaced by a priority queue. > Even using a queue, that’s a data structure that a person with a task to do isn’t likely to spend the time to invent on the spot. A to-do list is as simple as data structures get. A heap is complex, but if you want to insist on developing everything used by the algorithm from scratch, the person faced with this problem is likely to just use a sorted list. Try it out. Ask some people how they'd find the shortest path between two points through a weighted graph (use whatever terminology you want); see what they say.

zarathustreal 2y ago

Look brother you’re not winning any points by implying that you’d be the guy to invent A* on the spot. Without formal education you’re almost certainly not even dealing with concepts like breadth-first search, trees, or nodes, so no, a ton of people will not be inventing it on the spot. Edit: most people don’t even think about recursion as a possibility until taught about it in a formal setting, and even then it’s typically a struggle to understand. Even using a queue, that’s a data structure that a person with a task to do isn’t likely to spend the time to invent on the spot.

hnfong 2y ago
I agree the vast majority of people aren't going to invent A* on the spot. But I don't understand your emphasis on "formal" education. I'm pretty sure a sufficiently smart high school kid interested in the subject could learn all these topics online in a couple months without enrolling in any college.
thaumasiotes 2y ago
> Without formal education you’re almost certainly not even dealing with concepts like breadth-first search, trees, or nodes, so no, a ton of people will not be inventing it on the spot. If you're not dealing with nodes, you won't be inventing A* on the spot because you don't have a problem to which it can be applied. But otherwise, it's literally a breadth-first search with the queue replaced by a priority queue. > Even using a queue, that’s a data structure that a person with a task to do isn’t likely to spend the time to invent on the spot. A to-do list is as simple as data structures get. A heap is complex, but if you want to insist on developing everything used by the algorithm from scratch, the person faced with this problem is likely to just use a sorted list. Try it out. Ask some people how they'd find the shortest path between two points through a weighted graph (use whatever terminology you want); see what they say.