 | | Re: Mindstorms NXT programming languages —Juergen Stuber
|
| | | Hi William, (...) I'm pretty sure the official language also uses a VM like on RCX and Spybots, so there you have your "massive overhead". Lejos is a good approximation of a Java VM in the RCX, and the NXT has 2 to 10 times as much memory depending (...) (18 years ago, 8-Jan-06, to lugnet.robotics)
|
| | | |
| | | | | Re: Mindstorms NXT programming languages —Brian Davis
|
| | | | | (...) Actually, when using the standard firmware speed is very much a concern with the RCX for me... and others, if things like the Swan firmware are anything to go by. Currently, using NQC, I need to examine the bytecode generated to try to (...) (18 years ago, 8-Jan-06, to lugnet.robotics)
|
| | | | | |
| | | | | | | Re: Mindstorms NXT programming languages —Juergen Stuber
|
| | | | | | Hi Brian, (...) On the RCX Lejos is about that much faster than the standard firmware, and we have 1ms timer resolution. (...) You also get those in Lejos, and compared to NQC it can access more memory. Jürgen (18 years ago, 8-Jan-06, to lugnet.robotics)
|
| | | | | | |
| | | | | | | | NXT power (was memory, prog. langs etc) —dan miller
|
| | | | | | | My 2c based on some experience with high-level robotics programming (vision, navigation & mapping, etc): While it is concievable that you could write some sophisticated mapping and analytic software on this platform, it will require deep low-level (...) (18 years ago, 8-Jan-06, to lugnet.robotics)
|
| | | | | | | |
| | | | | | | | Re: Mindstorms NXT programming languages —Brian Davis
|
| | | | | | (...) Yes. And the Swan firmware is about that much faster, with a lot of added things as well, and I can still essentially use all my NQC code and tricks. BrickOS is much faster, etc. I guess my point was that I hope the NXT is faster (...) (18 years ago, 9-Jan-06, to lugnet.robotics)
|
| | | | | | |
| | | | | | | | Re: Mindstorms NXT programming languages —danny staple
|
| | | | | | (...) That would be a fantastic win - keep a very minimal kernel, and drive the software/hardware basics to be practically driven. Many languages could they be built on that -with an SDK that only then places the required language and tools on the (...) (18 years ago, 10-Jan-06, to lugnet.robotics)
|
| | | | | | |
| | | | | | | | Re: Mindstorms NXT programming languages —Ross Crawford
|
| | | | | | (...) Pretty small: (URL) (18 years ago, 11-Jan-06, to lugnet.robotics)
|
| | | | | | |
| | | | | Re: Mindstorms NXT programming languages —William Grant
|
| | | | (...) The bytecode interpreter on the RCX/Spybot is somewhat simpler than Java, I am pretty sure. (...) Lejos is a small Java VM, yes. As you say, whether a proper(ish) one is feasible depends on what access we have to the flash memory (64kb isn't (...) (18 years ago, 8-Jan-06, to lugnet.robotics)
|
| | | | |
| | | | | | Re: Mindstorms NXT programming languages —Juergen Stuber
|
| | | | Hi William (...) not much actually for the subset that Lejos implements. The most complex part of Lejos is the scheduler (multiple threads, waiting on monitors). W.r.t. speed we are faster, but the standard firmware has a more uniform timing. (...) (...) (18 years ago, 12-Jan-06, to lugnet.robotics)
|
| | | | |
