20 May
2013
20 May
'13
7:02 p.m.
Hi Paul,
Am 19.05.2013 11:51, schrieb Paul Birkel:
As a point of philosophy (or maybe practicality!) the
external "boat
anchor" is the first thing to get chopped up, lost, damaged, destroyed,
rusted-out, whatever. The controller on the legacy system is much more
survivable. In consequence we get the mismatch between "I have a
controller" and "I have a drive" (separate from the issue of "I have
a
*working* drive" ...). Emulating the controller-protocol (over modern
removable media) meets this real-world gap. And as Ethan points out, it's
portable across system-buses (*e.g*., PosiBus, Omnibus, Unibus, QBus).
One could
assume that. But... One controller connects to more than one drive.
And my RK8E experience is bad. I have two or three working RK8E controllers. And
around 15 or so RK05 drives.
Emulating the controller-on-the-bus avoids rewriting
system drivers, sure.
And this is what Philipp has done to great advantage in his recent work
:->. And I completely agree that this is a good thing. And I'd support
more of this good-thing if Philipp decides to do so for the drives that I
mentioned. It's just a different thing. And it's system-bus-specific.
And I think that many folks would prefer board-level removable media
(that's portable to a PC -- Linux and/or DOS/Win) rather than a PC-tether
:->.
To clarify again: It was not my primary intent to build anything that helps to
substitute vintage drives or controllers. I wanted a fast and reliable data path
between a pdp8 and a PC. Which helps at least me to backup and restore real
vintage media spinning on real hardware.
Then there's the rewrite-drivers for
new-hardware-with-arbitrary-protocol
approach. Not what we're talking 'bout here.
Some are actually talking
about that.
Apparently Philipp is blessed to have lots of
functional drives at his
"disposal" -- but maybe not controllers? Many/most of the rest of us are
perhaps in the boat of controllers-without-drives (and it appears,
varying-bus systems). And some have neither ...
See above.
I think that an economics-based case can be made for
the
portable-across-system-buses approach ("emulate the cable to the boat
anchor"). I'm surprised that there doesn't seem to be more work along
those lines. Maybe I'm just looking in the wrong place(s)?
What I would do to build such an emulation:
- Take a small computer - in my case probably beaglebone
- Do a board with interface facilities to the controller and a nice FPGA which
is connected there and to the little PC's bus
- Write the board emulation
This is not really a simple task. You have to emulate the drive with several
aspects of it's original behaviour. Then you have to work with raw sector data
(written before) which is not directly usable on a PC. And the conversion
between raw disk data and user data varies. It depends on the controller in use.
If the emulation is then running with one controller there is a quite good
chance that it does NOT with the next type of controller. Just because of subtle
differences between the emulation's and original drive's behaviour which matter
for one controller but not for another one.
For me this is more complex than building a single board PDP8 with lights and
switches on the backside of the PCB.
And I don't think that there are people who want to pay $$$$$ for such a
solution. Some might build their own at home. But that's it.
Kind regards
Philipp
--
Dipl.-Inf. (FH) Philipp Hachtmann
Buchdruck, Bleisatz, Spezialit?ten
Alemannstr. 21, D-30165 Hannover
Tel. 0511/3522222, Mobil 0171/2632239
Fax. 0511/3500439
hachti at hachti.de
www.tiegeldruck.de
UStdID DE 202668329