[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Uses for m68k Was: Re: debootstrapping m68k-coldfire

Brad Boyer wrote:

> > On Tue, Mar 04, 2008 at 09:17:54PM -0500, Michael Casadevall wrote:
> >   
>> >> I suspose the question needs to be asked; what are people doing with their 
>> >> old m68ks. Most people around here are using them for (obviously enough) 
>> >> buildds to attack the unstable queue. I popped over to netbsd, which is 
>> >> the only other operating system distro besides Debain and Gentoo to 
>> >> support m68k. The mailing list is completely dead looking at the archive, 
>> >> and the latest NetBSD releases do not offically support any m68k 
>> >> architecture as far as I can tell.
>> >>     
> >
> > I'll be quite honest about what I use my m68k machines for. They are
> > all turned off and mostly not even hooked up to anything unless I am
> > doing development on Linux. I use m68k as an excuse to look at things
> > like SCSI and ethernet drivers without any pressure to fix something
> > that people actually use. It's an excuse to mess around with stuff I
> > wouldn't have a chance to get code accepted on an architecture that
> > had real users. I don't expect to ever use any of my m68k systems
> > for any actual work. However, I would be a little disappointed if
> > there wasn't a distribution I could install and work on a current
> > kernel tree.
> >
> > 	Brad Boyer
> > 	flar@allandria.com
> >   
I guess I need to turn on popcon and be counted.

I have installed Debian on several m68k machines over the years (all
Macs, since my Amigas don't have the CPU or memory requirements, and I
haven't gotten my hands on a suitable Atari yet).  I haven't done too
much practical work with them to date, in part because there have always
been little software issues that have stood in my way.  But I believe,
from a philosophical point of view, that there ought to be an Open
Source operating system for the 68k machines that provides a comparable
level of functionality and software support to their original
proprietary systems.

I would actually use a 68k Mac for basic word processing with AbiWord or
Ted.  Just a couple years ago I wrote rather extensively using
WordPerfect on a Quadra 700.  Why shouldn't I be able to do the same
thing on the same machine using all Free / Open Source software?  I
would do light web browsing with Dillo, and possibly use Silpheed.  I
would also use netpbm to do some image manipulation -- possibly by means
of a web/CGI interface -- for things like converting images for display
on even older computers (Tandy Color Computer 3, MM/1, Amiga 500, 1000,
etc.)  I would even do a little programming in C, Perl, and Lua.  I
recently switched on a Quadra 630 with Debian so I could test tunneling
X over ssh in preparation for remotely administering a web server for a
client.  So I have actually used Debian m68k in support of a paying job
within the last week!

Sure, I could do all these things with other computers and other
architectures.  But the point is that I have the 68k hardware; it is (or
ought to be) capable of doing the tasks mentioned above, and a lot more;
and I want to see these computers become useful again, to the best of
their ability, and not throw them away, or pack them in boxes in the attic.

Why continue the Debian port to m68k?  Because existing Debian releases
have issues on m68k that warrant further development.  Perhaps the
direction Debian is going doesn't help matters, but I believe that some
kind of new OS development is needed for these machines -- if only a
final, polished farewell release -- and so far Debian has been the
closest thing to a usable Open Source OS for them that I have seen, but
it hasn't quite done yet what it should be capable of doing. 

I first installed Potato on a Mac IIci.  The main problem I recall at
the time (aside from the general slowness of the IIci) was that the
available software was pretty archaic (a perennial issue with Debian, I
suppose, but I noticed it most pointedly with Potato.)  What I wanted,
not because I need it so much for my own comfort, but more for the
principle of the thing, was a more complete, but still lightweight GUI
-- something that would replace MacOS functionally without the bloat of
GNOME or KDE.  Things like IceWM and Fluxbox fit the bill rather nicely,
and their continued development has remained focused on balancing low
resource use with functionality and ease of use.  More recent versions
of these and other packages provide more of what was lacking in previous
releases of Debian.

On the other hand, an unresolved issue I've had with more recent
releases of Debian is that (at least on the Mac) there is no key repeat
in X.  Console yes, X no.  I think it's an issue with xorg, and I don't
recall it being a problem on Potato (but that was a while ago now, and
my memory could be faulty.)  It doesn't happen when I run X apps
remotely on, e.g., an AMD K6-2 450 running Ubuntu.  It's just a problem
with the X server in m68k.  I believe I would actually use the old Macs
more (for the purposes stated above) if this issue were resolved. 
Perhaps I should try xfree instead of x.org.  But these are the kinds of
little glitches that a future release could address.  Older releases =
archaic software.  Recent releases = better software choices, but other
usability issues.

There is an unofficial patch for Dillo which adds support for frames. 
This version is used in Damn Small Linux.  I ran Dillo on Woody on a
Quadra 840av and was very pleased with its performance.  A few months
ago I made a first stab at getting the patched version to compile on a
Quadra 630 [Sarge, I think].  I ran into problems and had to set it
aside for a while, but I'm sure it can be made to work with a little
more persistence.  Now, the patched version isn't an official Debian
package, of course.  But everything else installed on that computer
needed to compile it is.  And the patched version is a significantly
more usable browser than the official one that comes with Debian.  The
point is that we've got something that is very close to being useful --
that could be much more useful than it is with a little, strategically
selected work.

Emile is very close to freeing us from legacy MacOS.  I installed Sarge
(IIRC) on a Q630 from an unofficial CD image that boots by means of
Emile.  IIUC, there's only a small bit of Apple code that needs to be
replaced in order to make a distributable, bootable Debian Net Install
CD image.  How many more people would at least try installing Debian on
an old Mac if such an easy option were available?  Maybe those Apple
drivers are too difficult to replace, but it seems to me that we're too
close to real usability to quit now.

Another reason that continued support for older, low-resource computers
is important is that they share many traits with handheld and portable
computers:  comparatively low memory, slow CPUs, small/low resolution
displays, low mass-storage capacity, etc.  Programs like Dillo,
Silpheed, SIAG Office, KDrive, and others can be useful on old desktop
systems and also newer handhelds, and the more places they can be
employed the more likely is their continued development.

Perhaps the m68k platforms would be better served by a smaller OS,
derived in large part from Debian, but not dependent on the Debian
organization or infrastructure.  But I have seen some of the prominent
members of this list express that they would not contribute to any m68k
effort that was not Debian.  With the loss of such expertise, I can't
see how any further development would be feasible.  But a smaller
release, with software hand-picked not only to run, but be potentially
useful on real 68k machines would be of more value to any potential end
users than a more comprehensive OS that is, for all intents and
purposes, unusable. 

What about the possibility of developing a Debian package ('task'?
suite?) designed specifically for installing a Debian subset on older,
more resource-constrained machines irrespective of architecture, with
arch-specific dependencies?  Then put up some explicit instructions for
how to use this to set up an old Mac or Amiga and see if we can get a
few more users/maintainers to justify continuing the port, while at the
same time shining a spotlight on some of the applications that are still
potentially useful on older hardware.

Again, I would use Debian m68k more if a few key usability issues were
cleared up, and I think more people would at least try it if we could
come up with a way to install it perhaps a little more easily on readily
available hardware, maybe with an option (in tasksel or something like
that?) for a pre-selected suite of lightweight apps, and if the
information on the Web could be kept a little more up to date.  How many
extra users could that garner?  Who knows?  But with as few users as
reportedly show up on popcon, it wouldn't take many to make a huge
percentage change.


Reply to: