[Coco] So, would you like a little Internet with your CoCo?

[Andrew]

I am following this by a little, and thought I would inject my own 
thoughts on this subject.

1) First off, I say port the uIP stack:

	http://www.sics.se/~adam/uip/index.php/Main_Page

    It's very small - according to the docs "the code size is on the 
order of a few kilobytes and RAM usage can be configured to be as low as 
a few hundred bytes".

2) Implement this stack so that it can be used under either OS-9 and 
DECB. It might have to be limited to OS-9 - the tricky thing is it is 
written in C, and so it would either have to be done as a cross-compile 
(isn't there a form of gcc to cross-compile to the 6809?), or it will 
need to be ported to Microware C - for DECB, porting it might be an 
almost impossible (or very long and hair pulling) task. I don't know if 
there is a DECB C compiler out there (maybe the object code could be 
pulled from what the Microware compiler creates).

3) The stack should be able to be set up to communicate with the network 
via one of the serial ports (bit banger or RS-232 cartridge), or over 
the DriveWire or Roger's Wireless pak. This setup should probably be 
done in such a manner so that the user can decide which to use; if it is 
made modular, then we can add extra devices in the future should they 
become available (ie, a real network interface card, for example).

4) It would likely have to be limited to use on a CoCo 3, or on a CoCo 2 
with 64/32k (smaller CoCo systems might be too small, but we might not 
really know the truth until the stack was implemented).

Once we had a working stack (with at least telnet working), then we 
could start writing other applications to communicate using it. As far 
as a web browser is concerned, I do believe that it needs to be a very 
cut-down system, the CoCo doesn't have the memory or speed to do 
anything too fancy.

We should probably stick to something like WAP, but even smaller - I am 
going to call it uHTML (Micro HTML: HTML pared down drastically). I 
would say the browser should be able to render the bare-bones basic tags 
needed for HTML presentation, maybe allow super simple forms capability, 
and maybe even a small amount of graphics (like 1-2 bpp images, no 
larger than 160x160, yield 3-6 kb image sizes). The data would come from 
a server set up to serve it; maybe someone could write an Apache module, 
or maybe one of the tiny open-source web servers could be modified to 
serve this special form of HTML (maybe someone has already did it for 
one of the other 8-bit micros?).

Lastly: regarding using a mini-ITX box to convert pages or do other 
stuff... Such a setup would be overkill. Yes, we are going to likely 
need a "pass-through" system for the CoCo because we don't have a real 
network card yet. But this pass-through system doesn't have to be 
complex, and it doesn't have to be standalone - the system could likely 
be any nearby PC. It just needs to interface its serial port with the 
CoCo's, and translate the communications over to a TCP/UDP port, which 
would then communicate with the rest of the network.

Now, of course, this is where we get into what could be done and how 
much you want this to be a CoCo thing. What I outlined above would make 
it mostly CoCo based. If you wanted to simplify the above, then 
implement the uIP stack on a microcontroller, and communicate with that, 
then have that communicate with the rest of the network via ethernet 
(plenty of easy to use microcontroller ethernet chipset interfaces out 
there). Or, implement the system on a BeagleBoard (its a very small 
embeddable system - almost like PC/104, but it has everything on-board, 
and it is really cheap), and stuff the BeagleBoard in the CoCo or in a 
cartridge.

I hope the above spurs more discussion and thought, and I wish I had the 
time, energy, money, and time to help implement it (don't we all!). 
Right now, other projects and things are taking priority, so all I can 
offer is words, unfortunately...

-- Andrew L. Ayers, Glendale, Arizona