Re: Drive speedtest PRG

On Sat, May 12, 2018 at 12:51 PM, Mia Magnusson <mia@plea.se> wrote:
> Den Fri, 11 May 2018 14:44:37 -0400 skrev Ethan Dicks
> <ethan.dicks@gmail.com>:
>> While I do think the M68K family is awesome, what could one do with a
>> 68008 in a CBM-II?
>>
>> It sounds theoretically cool, but what would it do?  What would be the
>> reason to put in that much effort?  What could it do that would give
>> it some "wow"?
>
> You ask a really good question. There aren't that much general 68k
> software that runs on anything else than specific hardware (Amiga,
> Atari ST, Apple Macintosh).

Right... the richest software ecosystems for the M68K require specific
video implementations at least (each of those platforms is non-generic
in specific ways).

> The only software I know of is CP/M 68k
> with afaik very few programs, OS/9 which should have at least some
> software

Those are good starting points.  I don't know how much stuff is out
there for CP/M 68k but there is definitely _something_ (i.e., not
starting from scratch).

> If the computer had some kind of bitmapped display hardware that could
> be programmed in a way at least slightly compatible with Amiga, Atari
> or Mac, it would be far more interesting.

The Mac has a very dumb video implementation, but the application side
requires the routines in the ROM Toolbox.

One thing to remember is that the 68008 has a smaller address bus than
the full 68000.  20 bits for the DIP verstion, 22 bits for the PLCC.
This matters when trying to emulate a memory map that expects 24
physical bits of address space.  One needs a little space for ROM and
for I/O carved out of the memory map, so 1MB of RAM is trivial to set
aside, 2MB isn't any harder.  3MB is a bit odd (no pun intended) but
barely any harder (if one uses GALs or CPLDs for implementing the
memory division).  There are other slightly more complex arrangements
(ROM that can be switched out after boot, small I/O window on top of a
sea of nearly 4MB of RAM, etc...) depending on how far one wants to
take things.  Minimum, I'd say, is quarter the map and do 1MB of RAM
space, 1MB of ROM space, 1MB of I/O space and ignore the remaining 1MB
space.  Viable but boring.

> Btw there were Unix-like operating systems for 68k computers too.
>
> But I guess that those unix-like operating systems needs a MMU.

There's one UNIX-like OS I know of off the top of my head that does
not require an MMU and that's Minix (I have the Minix disks for the
Amiga 500, but it would be a good starting point, after one replaces
the console interface from fiddling the Amiga custom chips back to a
serial line or an 80x25 character buffer (CBM-II video memory)).

> So a 68008 wouldn't be that useful. On the other hand, you could argue
> that everything we discuss on this list isn't that useful either, from
> a utilitarian perspective ;)

I entirely see the point with the recent project to get MS-DOS newer
than 1.25 running on the 8088 board.  It's taking an existing
functionality from back in the day and giving it a clear and useful
upgrade.  As for other widgets, we do discuss a lot of items that are
not exactly efficient or economically practical, I was kind of more
angling from the perspective of - OK... here's this new board.  With
no software it's an expensive lump.  It should do *something*.  One of
my personal goals is to run Zork on places that Zork has never run
before (I helped with the development of a Z-machine implementation
for the RCA CDP1802 a few years back).  Even to do that for this
theoretical board would take a bunch of groundwork - a cross assembler
at least.  (I've already gotten a Z-machine working on 40 and 80
column PETs as well as the VIC-20 (requires as much extra RAM as one
can stuff into it), but since I don't happen to own a CBM-II, I've not
gotten Zork running on one).

Picking a killer app or two would also advise the design process -
does it need graphics and sound or is a plain wall of characters and a
filesystem enough to justify the effort.

-ethan