Hi,

On Wed, May 2, 2012 at 7:46 PM, Jim Hall <jh...@freedos.org> wrote:
>
> I really like the core of Martin's idea, and I think it shows a lot of
> initiative and energy. I say go for it, give it a try, and see if it
> works. That's how these things start.

Sure, he can do whatever he wants. But it's easy to bite off more than
you can chew. More realistic is to pick a VERY narrow (but feasible)
short-term project that can be accomplished with only a few people.
Big things like drivers and OSes and heavy-duty things might be a bit
too impractical for now. But that doesn't mean you can't plan and
think about it, just make sure the idea is sound before beginning. (In
other words, user-space stuff is probably the easiest to do.)

> Years back, I had a somewhat similar idea, so maybe Martin can use
> this: Rather than writing any kernel stuff, or redefining existing
> APIs, recreate a Unix/Linux-like operating system that's based around
> a DOS kernel. You have the FreeDOS kernel to start with. Take the GNU
> utilities and create a Unix/Linux-like "userspace" of utilities. It's
> an interesting job of porting & recompiling to work on a different
> platform.

This is a great idea, and I agree, but it overlaps with DJGPP and
GNUish (somewhat). Doesn't mean there aren't plenty of gaps to fill,
but most people will probably not want to do this.

> A lot of this already exists in the "gnuish" utilities. Several of the
> FreeDOS utilities in the "Util" category are based off similar Unix
> utilities. What Unix utilities you cannot port to DOS, write
> replacements.

I've found Minix 2.x sources to be cleaner than others (not that I've
heavily looked), so things would be (relatively) "easy" to port from
there. Minix 2 was old POSIX.2 complaint, so should be easy, even if
not using DJGPP. (And OpenWatcom has some partial POSIX tools support
in their tarball.)

> I'd be really interested in seeing a feature-complete Unix/Linux-like
> command line system based on DOS.

Well, unlikely to 100% happen, mostly because POSIX is such a big ball
of wax, too many things going on. In particular, I hate the idea of
AutoTools as it's overkill, very brittle, barely works, and is (at
best) DOS unfriendly or (at worst) DOS hostile. We don't really want
to encourage lots of ugly *nix scripting relying on ten bazillion
POSIX tools. It's not that it's bad, per se, just gets out of hand and
too hard to fix or maintain (IMHO).

Take a look at this:

http://www.unix.org/version3/apis/cu.html

Even ignoring all the oddities (that we'll never need) and other
weirdnesses (pax, sccs), it's overkill.

But we do already have (unofficial) DOS ports (or very similar) for
various things (even totally ignoring DJGPP):  sed, awk, bc, cksum,
chmod, compress, date, vi, file, find, grep, iconv, ls, man, od,
printf, sort, strings, tee, touch, tr, uudecode, wc, ....

It doesn't mean more couldn't be done, polished, packaged up properly,
switched to free-r tools (on some) or slimmed down (on others). But
overall we're not that bad off. It's just not all bundled together,
necessarily. (Apparently most people are bad about keeping up to date
on things, heh. Linux has much more devoted fans.)

P.S. This may or may not make sense, but I think POSIX is showing its
age here: a bunch of (superceded) cmdline tools. Admittedly useful but
people flock towards bigger solutions these days (Perl, Python, Ruby)
or other languages entirely (other than C99/C++). In short, I've been
investigating a lot of languages lately (past year or two), still
*PLENTY* more to do in that regard!!    ;-)      So this isn't a bad
idea, but we'd need to isolate any specific areas that we want to work
on, first and foremost.

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
Freedos-devel mailing list
Freedos-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/freedos-devel

Reply via email to