At 08:35 PM 10/10/98 +0200, you wrote:
>I'm trying to develop a asynchone (or synchrone?) method to transfer data
>between 2 MSX-computers (i.e. for games), using JoyNet. Advantages of this
>compared to a fixed Baud-rate-protocol are that it runs on every kind of
>MSX-computer, unlike its speed and available timers. Also, you won't have to
>recalibrate each time after sending 10-20 bits. Both computers use the same
>speed as the slowest MSX, and the slowest MSX _will run on its top speed_.
My protocol (the one I sent you, and also sent to the list a while ago)
uses the same type of communication. It's called "asynchrone".
>The main idea is that bith computers flipflop bit 3.
Don't forget to mention that your game uses only 2 MSXes. Other JoyNet
programs will work on larger numbers of MSXes. The bidirectional transfer
you propose is only possible if you connect just 2 MSXes.
>Every transfer is a send + a recieve, both containing 2 bits (and 1
>flipflop).
Note that your scheme is not delay insensitive. If the clock (which you
call flipflop) arrives earlier than the 2 data bits, the old data can be
read instead of the new data, resulting in a transmission error. You do
have error detection, but retransmission is not as nice as avoiding errors.
Unfortunately, I don't think that a delay insensitive code exists that
encodes 2 data bits in a 3 bit signal.
>The first transfer transfers an ID-byte, indicating any send-requests plus
>the checksum. If both computers don't want to send, the transfer stops, if
>both computer wants to send, bidirectional transfer is used and if only one
>computer wants to send, the other computer does only read (and send
>flipflops).
In the case of one computer sending, you could also make the other computer
send garbage of equal length as the real data. That might make your
routines simpler.
>If you really want to make some
>network-programs, which will all have to be able to 'understand' eachothers
>transfers, then you might develop some standard PROTOCOL (which is not
>specified in the JoyNet standard.
Such a use would require a TSR, because even if one computer is currently
not engaged in a network transfer, it will have to forward messages from
other computers. That TSR could best be made as part of a driver.
But even then, it might be easier to standardize the driver interface and
still leave to protocol to the implementer of the driver. As long as all
MSXes on the network use the same driver, there is no problem.
And standardizing the driver interface allows other network hardware as
JoyNet as well. For example, the EPP cartridge Greg was working on.
>Ah, my homepage... almost forgot.
>The Official JoyNet homepage:
>http://datax.cjb.net/
Hey! Where did you get the "Official" qualification?
Bye,
Maarten
****
MSX Mailinglist. To unsubscribe, send an email to [EMAIL PROTECTED] and put
in the body (not subject) "unsubscribe msx [EMAIL PROTECTED]" (without the
quotes :-) Problems? contact [EMAIL PROTECTED] (www.stack.nl/~wiebe/mailinglist/)
****
