AW: CGA card (Mitsubishi Electric) with 192K RAM?

[Jules Richardson]

On 9/7/20 6:18 PM, Fred Cisin via cctalk wrote:
> Floppy boot seems like the next step.

OK, it boots off a DOS 3.3 floppy if that floppy is inserted before it 
attempts to boot from the hard disk. If I wait for it to do its "system 
file not found" bit, followed by a subsequent prompt to insert boot media 
and press a key, it attempts to access the floppy drive but then goes off 
into la-la land. Odd.

But anyway, taking the successful floppy boot route, I can certainly access 
the hard disk in terms of bringing up directory listings and TYPEing files 
to the display. So far, attempts to run anything from the drive just result 
in a lock-up (keyboard immediately unresponsive, hard reset required). 
There appear to be DOS utils on the drive, and command.com, but I've not 
checked for hidden system files yet. fdisk shows the partition as active.

> Got an IBM "Advanced Diagnostics" floppy to try?

No, but I see that the minuszerodegrees site has an image, so I'll write 
that out and see what happens.

Looking at the drive contents, incidentally, I didn't see anything that 
explains (or interacts with) that unusual video hardware - it basically 
just holds DOS and a bunch of documents written by the original owner. 
Maybe they got suckered into buying this fancy graphics hardware without 
having any actual need for it, and then of course EGA and VGA came along 
and rendered it obsolete anyway.

> XT controllers tended to NOT be interchangeable, even between various OEMs 
> of Xebec!

Yes - something that people often seem to forget, too. I've run into that 
quite often, where someone will hang onto an old drive because of the 
contents, but they'll dump the controller that it was formatted against.

> I don't know what the incompatability was.

I don't think there was any kind of standard at all for what the low level 
looked like - vendors were free to do what they wanted in terms of what 
values they used for flags and how they actually ordered things within the 
sector header. I suppose there were some tweaks made over time for 
optimization or reliability (or at least, recovery) reasons, too, which is 
why even a single vendor had a few different incompatible formats.

I expect it was the same in the SCSI and IDE worlds, but of course with 
those "the controller" which handles formatting is really part of the 
package, so it wasn't an issue.

Jules