 | | Re: Something else is needed, I think...
|
|
(...) That was my point, that it was not a benefit for having an interpreted "machine language" over using native machine language. The only benefits are portability and error checking, and the error checking is problematic when you get to that low (...) (28 years ago, 3-May-99, to lugnet.robotics)
|
| |
 | | Re: Something else is needed, I think...
|
|
(...) I guess it depends on what you want in your firmware. I have been working on my own object-oriented kernel, and I have written an object-oriented kernel for an embedded I/O controller before, and it is a big job. That is just the OS, not (...) (28 years ago, 3-May-99, to lugnet.robotics)
|
| |
 | | Re: Something else is needed, I think...
|
|
(...) This is good to know, I now know that any byte code interpreter that uses floating point will have to devote on the order of (at least) 6K to floating point if it wants to support floating point. Considering that any interpreter that uses the (...) (28 years ago, 3-May-99, to lugnet.robotics)
|
| |
 | | Re: Something else is needed, I think...
|
|
(...) Sorry to step in here. But there is no difference (as I see it) between having a system library function that runs the motors and having a byte code that runs the motors. It's all a matter of encoding. If I say: set_motor_speed(MOTOR1, 20); (...) (28 years ago, 3-May-99, to lugnet.robotics)
|
| |
 | | Re: Something else is needed, I think...
|
|
(...) Total .text section is 5832 bytes, .rodata is 4 bytes, and .bss is 8 bytes. This is for all for all of the FP emulation routines, only those actually used will be linked in. This is only the helper routines, not anything from the math library. (...) (28 years ago, 3-May-99, to lugnet.robotics)
|