22 Feb 17:00 2013

by Boise G. Pitre

Tandy Color Computer (CoCo) List

In all likelihood, Microware LP is not getting the source code to OS-9/6809 because it no longer exists as
far as anyone knows. The last known copy that I personally saw of the source code to the OS was around
'95-'96. At Microware, there was a large 6809 machine named 'Sybil' (Allen remembers this) which had an 8"
floppy and a 5MB hard drive. I remember Mark Hawkins pointing the machine out to me and indicating the
source might be there. I hooked up a dumb terminal to the com port, turned it on, and it booted into OS-9 Level
2 and gave me a shell. I staying late several nights at work digging into the file system and there it was...
all of the source code to the entire OS-9/6809 product. I could not get the source code off the hard drive at
the time because I couldn't find any 8" floppies
  to copy files onto.

Gosh, I even remember stumbling upon a text file that was a letter from Mark Hawkins to Kevin Darling about
trip arrangements to Des Moines to get the OS-9 Level 2 source code for the CoCo 3 for the Level 2 Upgrade.
That was a piece of CoCo history!

What happened to that machine is unclear. I called Mark Hawkins after leaving Microware in order to find out
what happened to it, and indicated that it was used for target practice (literally). Allen Huffman
indicated something else.

At any rate, it's gone, and from what I understand, that was the only place where the OS-9/6809 source code
resided. Microware did have tape backups of the 68K source and other software, but Microware LP isn't even
able to get that. From what I understand, they're only able to get all source code since the product was
managed in ClearCase (a source code management tool) from about OS-9 3.0 or 4.0 onward (mid 2000's vintage).

*IF* the source code to OS-9/6809 turned up, it would be quite unexpected and due to some old tape or hard
drive showing up that had it. In other words, quite by accident. The chances of that happening are probably
close to nil.

So what do we have for OS-9/6809 source? We have the NitrOS-9 Project, which has its roots in the NitrOS-9
product from Northern Xposure (Bill Nobel, Curtis Boyle, Wes Gale, and Alan DeKok), which has ITS roots in
a disassembly effort that Nobel/Boyle/Gale did years ago. That source has percolated up through the
years, became mixed with Alan DeKok's "TuneUp" product (which was a set of speedup patches for stock OS-9
Level Two by Tandy), and finally got some additional drivers and patches.

Then DriveWire sources were added, then SuperDriver sources (which I wrote) were added, and now you have
this huge amalgamation of code that is composed of:

1) Disassembly of original OS-9 Level Two modules
2) Added enhancements: 6309 instructions, patches, bug fixes and speed-ups like TuneUp
3) Drivers and booters for hardware that wasn't in the Tandy OS-9 Level Two product
4) DriveWire modules
5) SuperDriver modules

In short, we have a jambalaya of sorts...a mixture of code from a number of different sources not owned or
claimed by any one party per se.

Now, a lot of effort over the years went into making this big soup of code build and work on several computers,
including the CoCo. A whole set of tools (ToolShed) were built around this code to actually build it and
create disk images so that it could be deployed. Essentially an ecosystem has grown around it.

The NitrOS-9 Project represents the sole surviving remnant of the OS-9/6809 product which itself spanned
many systems (GIMIX, Tandy, Fujitsu, etc). NitrOS-9 is targeted to the CoCo, Dragon and Atari platforms,
all of which are defunct.

Since the 6809 isn't being sold anymore, and there is no perceivable market for OS-9/6809, it's very
unlikely that Microware LP would try to "take over" the project or demand its dissolution (which would be
nearly impossible anyway since the source code has been in the public domain now for many years; copies of
this stuff are on MANY hard drives out there). What Microware LP DOES have interest in and control of, is the
registered trademark name OS-9. The NitrOS-9 Project carries that name with impunity :)

This is no secret to anyone. The question has been asked before on this very list about the legitimacy of the
project given its pedigree. I chose for the most part to not answer those questions, knowing the thorny
issues involved. RadiSys, for whatever reason, chose to ignore NitrOS-9 completely, and there was no
need to wave the project in front of them to garner attention. So it was pretty much left alone.

But now with this acquisition, OS-9 is in the hands of a group of investors who actually know about the
hobbyist community and have some affinity for it. As I posted before... when the dust clears on the
acquisition, I will work with Microware LP to obtain some sort of legitimacy to the project and the use of
the name OS-9/6809. Remember, since they don't have the source code and in all likelihood won't get it,
this project represents their next best option. Not that they would ever be interested in selling
OS-9/6809 anyway.

What I will be proposing would be something along these lines:

1) The NitrOS-9 Project (or the OS-9/6809 Project if we rename it so) would exist as a hobbyist effort under
some open source license (Linville has volunteered to look into this in depth).
2) That effort would be recognized and sanctioned by Microware LP as such.
3) The name OS-9/6809 would be granted to the project for exclusive use.
4) Some recognition of the project on Microware LP's website.

I have every reason to believe Microware LP will be amenable to terms like this, with the following conditions:

1) Microware LP would not be responsible for the project, legally or otherwise.
2) The project's scope would be defined as supporting 6809/6309 processors ONLY in perpetuity. No other
processor would be supported.
3) The project would carry the proper name OS-9 without alteration (the 'Nitr' in NitrOS-9 would be dropped
and /6809 would be added, hence OS-9/6809 or OS-9/6309).

That's my take on the situation as it stands. We're wanting to release NitrOS-9 3.3, but with this monkey
wrench thrown into the works, I think it's best that we hold off until we have some agreement with Microware
LP, which could necessitate a name change and identity rebranding. Who knows, we might have something to
announce at this year's CoCoFEST at the end of April.

On Feb 21, 2013, at 5:22 PM, Allen Huffman <alsplace@...> wrote:

> On Feb 21, 2013, at 5:08 PM, John W. Linville <linville@...> wrote:
>> Wow, that is really cool!  I hope that works out!  Let's hope for
>> a nice, clean "open source" license for the 6809 code as well --
>> I'm happy to consult on the choice of a license... :-)
>
> One thing I think could be an opportunity for Microware LP would be having some sort of limited licensing
program for folks like us who would be interested in working with it and porting stuff over. I have been
learning stuff at my day job (CAN bus, SPI bus, etc.) that seem like they would be really easy to implement on
OS-9, but never were (though CAN bus came up from time to time). There is so much more reference code out
there these days -- and stuff that runs on little 16-bit Arduinos or TI MSP430s  in straight C -- like USB
stacks and the like. We might be able to contribute quite a bit if we had access to the OS and tools.
>
>                 -- Allen
>
>
> --
> Coco mailing list
> Coco@...
> http://five.pairlist.net/mailman/listinfo/coco

--
Coco mailing list
Coco@...
http://five.pairlist.net/mailman/listinfo/coco