[Coco] Re: atari USB device
James Dessart
james at skwirl.ca
Wed Nov 24 22:05:47 EST 2004
On 24-Nov-04, at 8:58 AM, Mark Marlette wrote:
> USB keyboard. Remember that what ever is developed should be backwards
> compatible with the existing software. NitrOS-9 is not such a big deal
> but RSDOS is where the problem lies. The new USB driver would have to
> be in ROM/FLASH at power up and patch in not and not to step on
> anything else. Then get the USB stream from the keyboard via the
> hardware device, whatever that maybe and then emulate the software
> call that originally was in the CoCo. POLCAT, peeks of the keyboard
> port, etc...If not, you will find that your keyboard won't work with
> certain programs after you have the CoCo booted up and you are able to
> type from the command line via the USB. Double edge sword here. That
> is why the hardware solution into the keyboard port was the choice at
> Cloud-9. It can be done, just a few layers to sort out, the joy of the
> beast. :)
I guess this underlines one of the big issues with any CoCo hardware
modification. Is RSDOS support wanted? Or are the people who want the
hardware modifications all OS-9 folks?
Would people expect to use a USB joystick with RSDOS games? Would a USB
mouse be used outside of OS-9?
Important questions, and it should drive any hobbyist development.
Personally, unless the device is for a single-purpose application, like
Contiki with ethernet on the CoCo (if it ever happens), then supporting
RSDOS is a pain. OS-9 is much better modularized, such that, in
general, it's easy to plug new devices in. But what would be even nicer
is to abstract the file system, so that regular NitrOS-9 operation
doesn't depend on RBF. In that case, though, there's a serious need for
a C compiler, because most of those who are most skilled with writing
abstraction layers are those who've had experience in higher level
languages.
Assembly might be fine for writing specific drivers, but for an
abstraction system you're looking at a lot of headaches. This is one of
the big reasons I wanted gcc to get up and going, because programming
with C is easier than programming with assembly for anything moderately
complicated. I've written mostly in C, but I do know assembly enough to
be able to say that the higher level the language, the more complexity
you can put into your programs easily. The more complexity, the more
flexibility and all the more features.
Just my $0.02 ;)
James
More information about the Coco
mailing list