Because this thread is going long and tracing it may be difficult for someone who wants to understand the core idea may read these first:
http://www.8052.com/forum/read.phtml?id=64218
http://www.8052.com/forum/read.phtml?id=64221
http://www.8052.com/forum/read.phtml?id=64268


First of all Erik,

I'm sure you are very professional and It's not like that I don't bother to reply to you, but as you see this thread has gone very long and gradually it becomes boring, many of the messages are unrelated to the core issue. I spend much time to explain the protocol ( or even if it is not a protocol in your opinion, whatever it is) and I think it's time to reach a technical conclusion, for example you may prove using fifos for async or buffered transmission is out of question or you may say it may work sometimes with some restrictions or.... I never intended to introduce a global IEEE like protocol, but it was just an idea that as I see, it will work at least for my current project, but you may refer to some technical pitfalls and change my mind. I reply to your messages but please forgive me (others too) if I don't respond to your term definitions and other non technical messages anymore and it's not because I don't bother myself, because then we miss the core issue. We all can open a dictionary and see what synchronous means, so discussing it here has no use and is just wasting of our times.


Erik said:
"¡and the moon is made of blue cheese.¡¨
You are right, not every protocol, but as I said I was pretty much preoccupied with USB. And usb defines both electrical characteristics and software layers. If I should apologize for that, I do. But as I said to andy I'm not going to argue about meanings of terms because then we never come to a conclusion about the core issue.

Erik Said, "demanding in which way?¡¨
Programming of it takes considerable memory, it is prone to error, it takes much processing time,...

Erik said "That would not help you much, one end must be a USB master and no '51 (yet) can do that.¨
I'm not sure I understood what is your mean, but from microcontroller in that specific sentence I didn't mean 8051..

Erik said
"¡§Using two FIFOs is hardly cheap¡¨ well in protocol that I've talked about I intended not to use supervisor chip so it may offset the price :)

and Patrick,

Patrick said
"¡§Please clarify above...do you have a single bidirectional 9bit wide data-path or two unidirectional 9bit-wide data paths?¡¨

http://www.8052.com/forum/read.phtml?id=64268 line 9

Patrick Said
¡§Please explain: What is a command I/O packet? is that a single 9-bit word with the 9th bit set to 1? How do you explain that "the remaining bits are crc for command + data"? You spoke about the 9th bit being set to 1. I then assume that the remaining 8 bits are the CRC. Where is the command?¡¨

It depends on how much instruction you want to define for example if you want to define 16 different instruction then you need 4 bits so remaining 4 bits are CRC.

Patrick Said
"¡§How long will you wait before you decide that something is wrong if the NACK or ACK does not arrive? How will you know what command has been ACK'd as obviously you do not want to wait until your command as been ack'd for sending the next command.¡¨

It may not be needed for all commands to be waited for ACK, but for those that need it depends on the situation of sender 8051, if the command is crucial to be acknowledged and sender 8051 cannot wait then it may send an interrupt as used in cross checking method and ask the other 8051 to complete it soon.

Patrick said ¡§What happens if you "miss" an ack? or more than one?¡¨
It should not happen, but if It happened for an unknown reason then Queued IPs variable gradually goes beyond MAX_ALLOWED_QUEDED_IPS and 8051 sends an interrupt and ask the other to complete IPs soon, this routine will not return until the other 8051 completes IPs or be reset and in all cases the variable will be reset to zero.

Patrick said ¡§see inline remarks above :-)¡¨

Yes I'm really open to technical criticisms, but please read above messages and tell me how many of them are technical criticisms about the issue.:)

Regards,
Ebrahim Vakilpour