Re: [MiNT] New MiNT distribution is n.. ..project

["George Crissman" <strads@tmisnet.com>]

>     About the hypothesis of a MiNT port for the forthcoming RioRed
> platform, I have some questions to ask : I think MiNT isn't a complete
OS,
> and is lacking some parts like the AES, the VDI, the
> desktop, and so on.

If I remember correctly, the "operating system" allows the computer
to communicate with peripherals such as parallel ports, serial ports,
MIDI ports, etc.  AES (Desktop) and VDI (Graphics) are part of an
environment that sites on top of the OS.  I'm still not too clear on
whether GEMDOS is part of the environment, although I suspect it is.
After all, BIOS / XBIOS handle floppy I/O as well as port communication.

> If MiNT is ported for the RioRed motherboard, those
> other parts should also be ported. What's the best way to procede ? To
> enter in a agreement with Woller&Woller, Behne, and so on, to port
> existing products like N.AES or NVDI ? To create a 680x0
> emulator for the PowerPC platform to use them with no port needed ? To
> create brand new AES/VDI/... parts ? I think the opportunity to port a
TOS
> get a TOS compatible PowerPC based platform is a *great* chance for us to
> perennize our world. 

If we're going to create a brand new AES/VDI, I certainly hope it would
add new calls and remain compatible with existing calls instead of
completely scrambling things up!  (Sorta like when "fsel_exinput()" was
developed to replaced "fsel_input()" -- the old call remained and the new
call used an unassigned ID number -- that's compatibility at work!!).  
A modular evnt_multi would be wonderful, as would direct calls to the 
joystick / joypad / mouse ports.  A new manager ("GemNet", maybe) 
providing calls to ease machine networking would also be nice.

> But I guess there will be problems on the way : a
> PCI/AGP platform means that sound and graphic output can vary. A hardware
> independant sound and graphic system has to be created for our platform.

The VDI is already an independant graphic system ... when fully
implemented, 
all calls may occur in a 32K-pixel x 32K-pixel workspace that is then
scaled 
to the actual display (say, 2048-pixel x 1024-pixel) in real time.  This
was not
reasonable when home computers ran at 8 MHz ... but the Milan / Medusa /
Phenix machines run SCADS faster!

As for sound, I would look at the World Wide Web Consortium (W3C)
and their emerging standards regarding "5.1" sound -- two left speakers
(front and back), two right speakers (front and back), a front center 
speaker, and a subwoofer.  Support for THAT would be visionary, and
would probably occur in the XBIOS, not in the environment.
W3C's web site is <http://www.w3c.org>, I think.

> and this is a huge work. I think Patrice Mandin started a project called
> DHI, which aimed to created such a software layer. Would it be a good
> starting point ? For the VDI, would it be possible to use Mesa, and to
> create a "gateway" to translate VDI calls into OpenGL calls ?
> Being a *poor* programmer, those question can be considered as pure
> nonsense from a hardcore coder point of view, so please don't flame...
:-)

Good ideas!

George Crissman
strads@tmisnet.com

Oh, yeah ... give me twin '060's, twin DSPs, and TOS/GEM/MiNT to tie it
all together... please ...

> 
> 
> Thanks !
> Yann
> 
> 
>