Yet another OS

Robert S. Thau (rst@ai.mit.edu)
Sat, 22 Aug 1998 10:20:35 -0400 (EDT)


Tim Byars writes:
> sheesh.
>
> http://www.freedows.org/english/main.html

You take this seriously? It's been kicking around for ages, but even
without that you can just look at the web site and see a few obvious
signs of trouble:

1) the group is putting a whole lot more effort into devising
"team structures" and other organizational matters, rather than
doing the minimum needed to support the core activity of
writing code. (Projects like this can go places if they have
a base of solid code to start working on --- but not from
scratch. The only way an open source project gets off the
ground is if someone, or at most a very small group, rolls
up their sleeves and resolves to ignore the politics and
get something done).

2) their technical emphasis is likewise... peculiar. The hard
part of writing a windows emulator, if you want it to do what
any sane person would want it to do (run microsoft apps) is
reverse engineering and implementing all the undocumented
APIs. Yet their web site soft-pedals that issue in favor
of talking up this academic notion of "cache kernels" ---
a new-flavor type of microkernel. Not that I think that
it's necessarily a bad idea (though if you're looking at new
ways to organize an OS, I'd favor MIT's exokernel stuff), but
it is at best a new way of solving problems for which there
are well-known solutions which are perfectly adequate.

rst