Re: NadaNet (was Re: brad w. Re: Ethernet & Internet on Apple 2)

[mjmahon@aol.com (Michael J. Mahon)]

Bill Garber wrote:

>"Michael J. Mahon" <mjmahon@aol.com> wrote in message
>20040703115957.25422.00000956@mb-m29.aol.com">news:20040703115957.25422.00000956@mb-m29.aol.com...

<snip>

>: My problem was in obtaining a unique temporary ID so that
>: the master could assign a permanent one through protocols.
>:
>: Many other network schemes use some unique address
>: built into the interface to prevent an endless loop of
>deterministic
>: collisions.  I hoped that I could use the (presumed) undefined
>: initial state of the video refresh generator to create the
>needed
>: non-determinism.  That assumption turned out to be wrong.
>:
>: Since I was already using Paddle 3 as a pulse-stretcher for
>: a "send" LED on each board, I have now varied the paddle
>: resistor value just enough to guarantee unique paddle 3 read
>: values, which the boot ROM now uses to generate its temporary
>: ID.  Problem solved--but I had hoped to do all boards
>identically.
>:
>: Because of the rather large tolerance on the paddle timing
>: capacitor, it is necessary to keep a record of the Paddle 3
>: value for each board, and to install a resistor that produces
>: a unique value.  ;-(
>:
>: A less elegant solution, but it works well with six boards.
>;-)
>
>Now a bigger problem exists. You've created a system which
>requires somewhat of a technician to install and set up. :o)

This problem certainly applies to the AppleCrate, which
would require some construction ability in any case.  There
are other solutions to assigning unique IDs for a NadaNet
composed of machines with keyboards, such as manual
entry, or using the keyboard-based "random number" counter
supported by the Monitor.  Typically, such machines will have
disk drives, and would not even require network booting.

That being said, I confess that I have not been too concerned
about "ease of use" issues other than my own at present.  My
main concerns have been the adaptability of the network to a
fairly wide range of Apple II platforms without add-on hardware,
the ability to construct networks of up to 16 machines, and the
usable speed of the network.  Simplicity of design has also
played a major role.  I expect that most ease of use issues
will be addressed at the application level, given proper enabling
at the lower levels.

I did hope that there was a common mechanism for generating
a non-deterministic value on an Apple II, both at Reset and at
power-on.  I had hoped--and led to believe by code in the self-
test ROM--that VBL might provide such a mechanism, but it
did not.

If anyone knows of such a mechanism (based on a really
indeterminate initial state), I would be happy to hear of it!

In the mean time, having a paddle resistor experimentally
chosen to produce a non-overlapping value with any of the
other seven boards does the job, requires no additional
components beyond my original design, and is marginally less
annoying than having to burn unique IDs into the boot ROMs.

-michael

Check out amazing quality sound for 8-bit Apples on my
Home page:  http://members.aol.com/MJMahon/