Re: Apple][GS-X (was: Apple II just won't die)

[sheldon@netcom.com (Sheldon Simms)]

In article <Pine.LNX.3.91.960317184926.31317A-100000-100000@cheetah.spots.ab.ca> "Brian D. Uhreen" <bduhreen@spots.ab.ca> writes:
>	
>	So far there have been numerous postings about the 'Rebirth' of
>the Apple IIGS in the form of a theoretical ROM 05 or GS-X. 
>
>So far the current 'Standard' that is presented.  If its in '[ ]' its my 
>addition:
>
>1)  Processor, a 65C832 or 65C1632, at a 40+ Mhz (you'll see why further 
>down around 7).  [This will incorporate the reguelar 'onchip' cache from 
>WDC as well as a possible total of 4 Megs Level 2 cache.]
>
>2)  Built in periperals to allow conections to previous AppleII devices 
>(ex. Second Sight, SWIM chip)
>
>3)  On board EIDE controller, is possible and debatable, this allows 
>access to the cheep PC Hard Drive market however a SCSI 2 wide 
>controler would be a more respectable standard in that its faster, and 
>less processor dependant.
>
>4)  PCI bus, [ and ISA, NUbus??]
>
>[5)  Sound, The 'S' in GS was put there for sound.  Lets not degrade 
>ourselves to the IBM type sounds (16bit FM synth).  Just staying with 
>ensoniq I suggest the OTTOFX an Integrated Synthesizer/DSP Chip.  This 
>chip can do:
>	  Wavetable synthesis with 32 independent voices
>	  Multimode 4-pole digital filter for each voice
>     	  On-chip memory manager allows OTTO, ESP, and CPU to share and 
>access memory without collision
>      	  Separate CPU, DRAM, and Sound ROM bus connections
>	  Parallel architecture - MAC, ALU, and address generator ALU
>This is one of the BEST sound chip on the the market today.  It could be 
>put either on a specific daughterboard slot or in an 'expandable' section 
>of the motherboard (more DOC RAM, chip swap out etc...).  No I don't work 
>for this company.]
>
>6)  Built in ethernet [100Base T]
>
>[7) Multiprocessing.  Seperate processors (MAC, IBM, PowerPC, etc) on
>seperate daughter cards with the ability to share the main memory and
>peripherals with the GS.  This could be integrated by treating the other
>processors as objects and allocating memory, resorces, and privilages as
>needed.  The GS would remain in charge and would take the appropriate
>code, send it to the correct processor (if installed).  Also each
>processor would have its own bit of meomory where the OS of that processor
>resides.]
>
>8)  The GSOS (insert cool spash screen here) 7.0.  This would take 
>advantage of the native 32bit code of the 65C832.  And also finally fix 
>all the bugs that we've been complaining about :-)
>
>[9)  The idea of OS's residing in flash ROM seems like a good idea to 
>me.  This will make a fast and easy storage medium for an operating system 
>and toolboxes and allows for patching.]
>
>Let me know what you think (remember this has to last for at least another 10 
>years, lets make it good ppl :-)
>
>	Brian
>
>
>  To thine own self be true.  And it must follow, as the night the day. 
>	Thou canst not be false to any man -- William Shakespeare
>		Brian Uhreen,	bduhreen@spots.ab.ca
>
>

Newsgroups: comp.sys.apple2
Subject: Re: Apple][GS-X (was: Apple II just won't die)
Summary: 
Expires: 
References: <Pine.LNX.3.91.960317184926.31317A-100000-100000@cheetah.spots.ab.ca>
Sender: 
Followup-To: 
Distribution: 
Organization: Netcom Online Communications Services (408-241-9760 login: guest)
Keywords: 

In article <Pine.LNX.3.91.960317184926.31317A-100000-100000@cheetah.spots.ab.ca> "Brian D. Uhreen" <bduhreen@spots.ab.ca> writes:
>	
>	So far there have been numerous postings about the 'Rebirth' of
>the Apple IIGS in the form of a theoretical ROM 05 or GS-X. 
>
>So far the current 'Standard' that is presented.  If its in '[ ]' its my 
>addition:
 
Interestingly enough I had been thinking about this for about a month
before this thread started so I was very interested to see what people
had to say...

>1)  Processor, a 65C832 or 65C1632, at a 40+ Mhz (you'll see why further 
>down around 7).  [This will incorporate the reguelar 'onchip' cache from 
>WDC as well as a possible total of 4 Megs Level 2 cache.]
 
I there such a chip availiable from WDC or are we really talking
fantasyland here? Will such a chip ever become available? The old
description of the 65832 (in Eyes & Lichty for example) was that it
would be pin-compatible with the 65816 - not very desirable in my
opinion.
 
My candidate for the main CPU is the 200Mhz StrongARM 110 from Digital
Equipment. I think that a very nice 65816 emulation would be possbile
especially if there is some hardware help. Alternatively, the 14Mhz
65816 parts are cheap enough that you could include both CPUs. That
brings up another good point for the SA-110, it's cheap. I had a paper
or article (that of course I can't find right now) that listed the price,
if I recall correctly it was in the $50-$75 range.

>2)  Built in periperals to allow conections to previous AppleII devices 
>(ex. Second Sight, SWIM chip)
 
Well a SWIM equivalent would be very nice, but I'm not so sure about
Second Sight - it depends on whether any significant software appears
that uses it. There's also the question of compatibility of serial
ports, ADB, etc. I'd be inclined to have this compatibility.

>3)  On board EIDE controller, is possible and debatable, this allows 
>access to the cheep PC Hard Drive market however a SCSI 2 wide 
>controler would be a more respectable standard in that its faster, and 
>less processor dependant.
 
This is a tough one. I think sticking to a PCI card might be the best
alternative. That's approximately the current situation in the IIgs
so you wouldn't be losing compatibility.

>4)  PCI bus, [ and ISA, NUbus??]
 
I think a PCI bus would be an absolute must have. ISA and Nubus are
worthless in my opinion. I do think it should have a few (maybe 3)
regular 'Apple II bus' slots, so people could preserve hardware
and also to provide early capability (before drivers have been
written for generic PCI cards for example).

>[5)  Sound, The 'S' in GS was put there for sound.  Lets not degrade 
>ourselves to the IBM type sounds (16bit FM synth).  Just staying with 
>ensoniq I suggest the OTTOFX an Integrated Synthesizer/DSP Chip.
 
This was my choice also, but I tried to get technical information from
Ensoniq and was told that the chip is proprietary. What a joke, so
is the PowerPC but if I sneeze in Motorola's direction I get a user's
manual tossed back at me.

>This is one of the BEST sound chip on the the market today.  It could be 
>put either on a specific daughterboard slot or in an 'expandable' section 
>of the motherboard (more DOC RAM, chip swap out etc...).  No I don't work 
>for this company.]

I think it can be on the motherboard, but is should definately have
expandable sound RAM mapped into the CPU's address space, with a
GS compatible access port for the low 64k of course.

>6)  Built in ethernet [100Base T]

Why not just use a PCI card?

>[7) Multiprocessing.  Seperate processors (MAC, IBM, PowerPC, etc) on
>seperate daughter cards with the ability to share the main memory and
>peripherals with the GS.  This could be integrated by treating the other
>processors as objects and allocating memory, resorces, and privilages as
>needed.

I don't know. I think that if I wanted a PC or Mac I would just buy one.
I don't see any reason to add special support for emulator boards.
I do think the ability to add other processors to increase computing
power is a good one, but it would probably be a lot easier to confine
extra processors to be the same kind as the main CPU.

>8)  The GSOS (insert cool spash screen here) 7.0.  This would take 
>advantage of the native 32bit code of the 65C832.  And also finally fix 
>all the bugs that we've been complaining about :-)

This would be something more than GSOS: a fully preemptive multitasking,
memory protected operating system. I do think it would be good to make
it very much like GSOS - maybe even preserving some source code
compatibility (or close to it at least). It is important though, that
the computer be able to run the current GSOS system 6 at the beginning.
I see the computer as being a hardware compatible system that runs
standard GS software and uses GS peripherals at the beginning then,
as the software is improved - new OS written, drivers for PCI cards and
peripherals, etc it becomes much more. Kind of the way early IIgses spent
their time running P8 programs before the IIgs stuff appeared.

>[9)  The idea of OS's residing in flash ROM seems like a good idea to 
>me.  This will make a fast and easy storage medium for an operating system 
>and toolboxes and allows for patching.]

Definitely. Flash memory is incredibly handy. It is possible for it
to be accidently wiped, so you would have to have at least some kind
of loader in a real ROM, that could reload the flash if the unthinkable
happened, but that could probably be very small, with the majority of
everything in flash.

>Let me know what you think (remember this has to last for at least another 10 
>years, lets make it good ppl :-)

I was thinking a little about video modes too. Because of the historical
legacy, the Apple IIgs video modes are outrageously complicated. I think
that perhaps 8-bit apple II support should be handled entirely through
software, and that therefore everything before superhires would not be
a hardware feature. Even superhires probably could be done *mostly* in
software, with some hardware help. I think the system should have a
standard built-in graphics 1024x768 true-color graphics mode since just
about any old monitor can handle that. It would be great to have an
even better mode built in, but I don't think it would be good to force
people to buy a $600 monitor just to use the computer.

In any case, the built in graphics capability can't be just plain
1024x768x24. some kind of indexed color capability would be nice for
palette effects, multiple graphics planes, maybe some built-in
hardware graphics engines, but it can't be expensive - the system
has to be affordable - we're not trying to compete with SGI. I also
like the idea of more video interrupts: interrupt on a pixel for example,
rather than just on a scan line. This area need some imagination.

One problem is that you can think of so much stuff, but in my
opinion cost is an overriding issue. If the result costs $2000 just for
the box, there won't be that many people shelling out. I think that
the cost of the basic computer (not including peripherals, monitor, etc)
should be $1000 or less. That's why expandability is very important.
Building in potential capabilities is better than having the basic system
ship with 8 megs of wavetable RAM, 4 processors, built in stereo speakers,
etc.

Also someone ought to dig up Todd Whitesel's old "Apple IIf" paper. It
was written a while ago when the idea was to have the Apple II compete
with the Amiga, but I remember it was interesting. I'll look for it after
I post this.

-Sheldon