[TI-H] Re: TI Networking.


[Prev][Next][Index][Thread]

[TI-H] Re: TI Networking.




At 22:12 2001-05-02, you wrote:
>One bit is sent like this:
>
>Sending calc brings red line low
>Recieving calc brings white line low
>Sending calc brings red line high
>Recieving calc brings white line high

hmm.. interesting. I have to try that. 4 lines to change in my pic program 
too see if it still works :)
I thought the sender handled both lines, that is what I have read, and had 
working when I used it. ofcourse, If I _also_ bring the other line low, it 
wouldn't change the function, else then that I do unnessesary work.
If it is like this, then we can't use the bytesender on the 89, too bad, 
but not an extreme loss..
I will research it.


>I think variable length might be good if you could allot less that a
>whole byte to indicating length, but that can be hard to do.  You could
>have an end packet marker rather than a start packet marker and use that
>to determine packet length as well.

mmm, or both a start and an end and infinitly long packets :)
a lot of options. I guess hands on experimenting is best to decide.

>You could have it turn based, each calc takes turns being the "master".
>Each calc would need to know how many others are out there, then every n
>packets sent, it sends one (whether or not it has anything to send).
>Similar to token ring, but on a shared medium.

that is a good idea. the problem is synchronization. who will start to 
send? who will be next?
this could ofcourse be solved with always filling up the adress space from 
the bottom, and let them send in order..  also have a "I have nothing to 
send" packet.

///Olle




References: