On Wed, 21 May 1997 13:09:30 -0400 Andy Kunz <montana@FAST.NET> writes:

>Harold,
>
>That was a pretty ingenious token system.  Not bad!

        Thanks!  I acutally called it "absence of data token passing"
when I first came up with it.  Since then, I've found (I think) that it's
called "minislotted".  In a "slotted" system, a fixed amount of time is
available for each site to transmit, whether it has data or not.  In a
minislotted system the fixed slot is used to indicate an intention to
transmit data, then it is expanded as necessary to transmit the data.
This gives "bandwidth on demand" that a slotted system doesn't do.
        My "absence of data token passing" idea was an adaptation of
typical token passing, but using the lack of data as a token to avoid the
possibility of a token getting lost and having to reconfigure the system.

        I agree that shifted checksum might be considered a simplistic
CRC.  I was never able to completely figure out CRC, so I invented my
own.  Standards often seem to try to anticipate every possibility,
burying us in details.  Because of that (and my not being able to figure
out the standard!), I invent my own.  Not necessarily the best way to do
stuff... but it gets it done.

Harold