| |
In lugnet.robotics.rcx.legos, Kekoa Proudfoot writes:
> You will miss about 1 in 256 errors that you do not catch earlier using
> collision detection. I think your error rate per message will be much less
> than 1 in 256 if you (can?) get the collision detection code right.
>
> But we shouldn't argue the point. Somebody should just code it and try
> it. Practice will show whether a CRC8 is sufficient or whether a CRC16
> will be required.
Hi,
Didn't mean to come across as arguing - sorry if I did!
As you say, lets get coding LNP (My assumption is that the coding bit hasn't
started yet, if it has, please accept my apologies & let me know how it is
going!).
I guess the things that need putting together are:
1) Line level protocol (Basic frame rx/tx with collision detection, CRC work
etc.). Has provision to send messages on request, and passes rx'ed messages
with good CRC etc. up to higher level. (i.e. it performs the basic MAC level).
2) Higher level of protocol. Does retry / acknowledgement / whatever. (Performs
kinda dumbed-down TCP like function).
1 & 2, if possible, written in OS independant C.
3) WIN32 support for the above. Provides basic character TX/RX, timers &
whatever.
4) Ditto for Linux?
5) Ditto for MacOS (Or is that McWindows, to annoy 2 companies whose lawyers
dwarf the TLG lawyers).
6) Oh yes, maybe something for LegOS... Maybe basic support + 1 & 2 'As-is',
or 1 & 2 re-engineered to fit LegOS more efficiently.
In addition some IP<->LNP router code would be nice.
1st thing to do would be to decide the interfaces between the above blobs &
then dish them out.
I'm up for taking some of the code in 1, 2, 3 & 6. I'm also keen to put
together a simpler 1 & 2 to get something working (maybe just PC<->single RCX
only) - if we do it right the final implementation could be dropped in its
place.
Thoughts????
Kevin.
|
|
Message has 1 Reply:
 | | Re: LNP - lets get coding.
|
| (...) No, you didn't - I just didn't wanted to make sure I wasn't :) (...) This has already been done, for Librcx at least; don't know if anybody took a look at the code I put at: (URL) mentioned wanting to maybe take a look at that; porting to (...) (25 years ago, 31-Jul-99, to lugnet.robotics.rcx.legos)
|
Message is in Reply To:
| | Re: LNP Repost (Me banging on about checksums)
|
| (...) You will miss about 1 in 256 errors that you do not catch earlier using collision detection. I think your error rate per message will be much less than 1 in 256 if you (can?) get the collision detection code right. But we shouldn't argue the (...) (25 years ago, 29-Jul-99, to lugnet.robotics.rcx.legos)
|
21 Messages in This Thread:
 
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
This Message and its Replies on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
