You can consider all those variables you mentioned, but it takes more memory and processing cycles to do so.
The benefit of an automatic go-left rule, a standard in the same way USB, the keyboard layout, and the alphabet song are also standards, is that it doesn’t take computational resources to decide on a solution. As long as the standard describes a working system, the inefficiency of its match to the myriad situations in terms of execution, is made up for in terms of efficiency of the process of deciding on that solution for that given situation.
USB as standard pays for itself by allowing us to buy computers and peripherals without having to think about connectors. Both of those things can be designed without effort spent on those decisions.
Always going the same way works well enough in almost all 2D navigation contexts that it creates a reliable way to avoid collisions at a societal level. In the ethernet standard , where there is no right or left because it’s a 1D context, they’re forced into a different anti-collision strategy: wait a random amount of time then try again to send. It’s less efficient than “take one step left, then proceed”.
I’m rambling. I hope you get the point. Standards save computational resources by loosely fitting a simple line to a complex data set, but in design space. There’s error, but it’s within acceptable range and has benefits in a different dimension.
Consider computational efficiency though.
You can consider all those variables you mentioned, but it takes more memory and processing cycles to do so.
The benefit of an automatic go-left rule, a standard in the same way USB, the keyboard layout, and the alphabet song are also standards, is that it doesn’t take computational resources to decide on a solution. As long as the standard describes a working system, the inefficiency of its match to the myriad situations in terms of execution, is made up for in terms of efficiency of the process of deciding on that solution for that given situation.
USB as standard pays for itself by allowing us to buy computers and peripherals without having to think about connectors. Both of those things can be designed without effort spent on those decisions.
Always going the same way works well enough in almost all 2D navigation contexts that it creates a reliable way to avoid collisions at a societal level. In the ethernet standard , where there is no right or left because it’s a 1D context, they’re forced into a different anti-collision strategy: wait a random amount of time then try again to send. It’s less efficient than “take one step left, then proceed”.
I’m rambling. I hope you get the point. Standards save computational resources by loosely fitting a simple line to a complex data set, but in design space. There’s error, but it’s within acceptable range and has benefits in a different dimension.