GILLMANN@USC-ISIB.ARPA (12/05/85)
From: Richard Gillmann N6MPP <GILLMANN@USC-ISIB.ARPA>
C2974 CC105 Lyle Johnson (WA7GXD,2973) 11/30/85 8:23 PM L:139
KEYS:/NNC SOFTWARE DEVELOPMENT/PLEASE SPREAD THE WORD/HELP WANTED/
To: All Interested Packeteers
Fm: TAPR
Re: NNC Development and Testing
As many of you are already aware, the new TAPR Networking Node
Controller (NNC) is nearing completion of prototype hardware
debugging. The NNC is a four-port packet controller with
a large memory area, direct-memory access (DMA) capability for
I/O -- and Z-80 software compatibility.
To dispell any rumors, there presently exists NO SOFTWARE for this
device. None. Nada. Ayn. Zip. Zero. Effes. Klum. { }
What we are looking for are volunteers to assist in developing
software for this device.
We need low-level, highly-efficient drivers for the I/O. We need
an AX.25 Level Two handler that can handle multiple logical and
physical channels. We need Level Three and Level Four. We need
loaders for uploading software updates to a remotely-sited NNC.
We are hoping that there will be early porting of multi-port
digipeater code to this unit as well as a W0RLI PBBS. We need
close coordination of the various aspects of the development. We
need... You get the idea.
The hardware should be verified during December. If all goes well
(it usually doesn't), we will want to put Alpha units in the hands
of developers in late January/early February. Assuming a couple of
months to get enough software together to make Beta testing meaningful,
we will be looking for Beta testers in the March-April timeframe.
Once testing has advanced to the point of reasonable confidence, we
will make the units generally available (summer of 1986?).
Now, we are NOT looking for folks who want to be the first kid on the
block with a new toy. We need people who are committed to Amateur
packet radio and want to help make a meaningful contribution to a very
large and difficult task.
And be forewarned. You may slave away for many, many hours, only to have
your code not used, or superceded, or... No guarantees.
Coordination is going to be a tough assignment. Without proper coordination,
a lot of wheels will spin, and a lot of energy wasted in duplication of
efforts. A BBS to swap code modules will be needed. All code will need
to be carefully, accurately and exhaustively documented -- by the author!
Developers will need to procure the following:
1) One NNC digital unit - projected cost is $175. This is an NNC with
uP, 64k bytes of bbRAM, 32k bytes of EPROM, four HDLC ports, two
parallel (centronics compatible) ports, two async ports and one
SCSI interface. The SCSI chip may not be included at this price,
we are not sure yet, but for the Alpha testers/developers it will be.
This unit will be fully assembled and "tested."
2) One NNC Floppy Adapter - projected cost $125. This includes a DMA'ed
Floppy Controller that can handle 4 diskette drives. This unit
will NOT support 8" drives (lack of 8" support is intentional). The
price includes a licensed copy of Z-DOS, a CP/M 2.2 compatible operating
system. It will be on 5.25" double-sided 48 tpi diskette format capable
of 368k bytes (formatted). If the decision is made up front to use 96
tpi drives, TAPR will copy the licensed diskette to the denser format
and supply the original as well as the high-density copy to the purchaser.
3) A pair of 5.25" floppy drives. Maybe we can do a group purchase of
TEAC 55Bs. Figure $150 for this expense. 48tpi or 96tpi are about the
same price. 48 tpi yield about 400k formatted bytes; 96 tpi about
800k formatted bytes.
4) A power supply. $50 from surplus sam?
5) One NNC Modem board - projected cost is $150. This is a wired and
tested board which includes one 300-baud 2206/2211 modem with
tuning indicator and three 1200-baud 2206/2211 modems. We might get
this cost down to $125.
Thus, there is a cost of participation that will be a minimum of $450
and may be $650. Add to this the cost of an assembler or compiler...
The assembler that seems to make the most sense is ZAS, from Echelon
systems. Again, we can probably do a group purchase or multiple-site
license for this project. This assembler supports the extended instruction
set of the HD64180 cpu. There is no reason to limit ourselves to the
Z80 instruction set (or - yeech - the 8080 subset) for this project. And
ZAS is fairly cheap - about $50, I think.
I don't know which C or Pascal compiler will be chosen. I assume that
one will be chosen, so the high-level code can be written in a transportable
high-level language (makes for easier testing?) while the interfaces to the
hardware can be done in assembly language. Preferably, the compiler will
generate Z80 (or 64180) source code for assembly by ZAS. This allows
hand-optimization of the compiler output.
By standardizing on the development environment (NNC w/5.25" floppies)
and the tools (assembler, compiler(s)), we hope to make it easier for
all participants to share their work amongst the group.
It is expected that all code (including source code) developed for this
project will be placed in the public domain for non-commercial use. And
that TAPR will be given explicit (not exclusive) right to distribute it.
If you have the time and ability and want the chance to make a real
contribution to Amateur packet radio networking development, please
leave a message on DRNET or write the TAPR office. We will put
you on file and notify you when we are ready to get started with Alpha
test or Beta test (as you indicate to us).
For Alpha test, we need developers. Committed developers. People who
really understand software design, hardware/software interaction,
protocol implementation, code size/speed tradeoffs, data structures
and myriad other facets of software design. And of course, understand
networking...
For Beta test we need testers. People who are in a real packet
environment, who have a good site that will get plenty of exercise
on the air, who have the time and commitment to submit detailed
reports of what works and what doesn't. This isn't a "be the first
person on your block to own an NNC" contest; it is going to require work.
What do we mean by commitment?
Consider a plate of bacon and eggs. The chicken was actively involved;
the pig was committed.
If you are a capable packeteer, committed to assist in networking
development, please provide us with the following:
Full Name.
Amateur Callsign.
Mailing Address.
Daytime telephone number.
Evening telephone number.
Alpha or Beta test.
TAPR membership number (if applicable).
Specific areas of expertise that you wish to make available to
this project (low level interface/high level protocol implementation/
documentation/testing/etc.)
Thank you for you help. Happy packeting!
[From the WB6YMH-2 mailbox.]
-------