Because the To: line contains many names, please be careful if/when
replying.
For very little money (I figure starting around $200) anyone can put together
a PC that is able to boot a unixish kernel from a floppy disk.
Such a PC might have no other storage device than the FDD (no HDD) and
just a simple VGA with monochrome monitor or RF adapter for a display.
Actually, I shouldn't be so unix-biased. Obviously DOS can be booted
from a floppy as well and my understanding is that there is a free
version of DOS available and considerable amounts of DOS-based Free
Software. DOS has some advantages over unix for some applications because
it is much smaller, leaving more room for the application itself. DOS
probably boots faster, too. It has disadvantages because it does less
than Unix. For example, is there a Free TCP/IP stack for DOS?
You can't fit a complete, traditional unix root file system on a floppy, at
least not easily, but you can easily fit a root file system in which
the "init" program is a stand-alone application. It might be a game
program, or a text editor -- or a language interpreter like Systas Scheme.
A Scheme interpreter is a good choice for "init". It fits on the floppy
and leaves room for some Scheme programs. Scheme programs tend to have
good code density, so you can fit a lot of useful functionality on one
floppy that boots unix, copies itself to a ramdisk, and runs Scheme.
For just a few hundred bucks, you can have a powerful lisp machine! Not
quite powerful enough to be self-hosting -- a machine big enough to
handle GCC is needed for that -- but definately big enough to run a
Scheme programming environment and/or simple applications like a small
text editor, an email program, a spreadsheet, etc.
Many programs and program fragments already exist which can be snapped
together to form interesting applications for such inexpensive machines.
This is a very rich playing field.
If you doubt these machines are useful, compare their compute power and
functionality to the microcomputers of the 70s, which did less and cost
an order of magnitude more money. Using free software and inexpensive
hardware, we have hit an unprecedented "price point". We're making
computing accessible to more people than ever before.
Personally, I have many ideas for how such machines can be put to good
use. I'm sure all of you can come up with similar ideas and more.
If any of you are near Berkeley, please consider visiting The People's Park
around 3PM tomorrow, 10-Apr-1997 to discuss some of the implications and
make plans relating to this.
-t