I received this reply from http://groups.yahoo.com/group/freeos
"Just to avoid the feeling that you are
speaking to a black hole
I will offer something of a response. I
think the general
answer is "yes". We could benefit from
a number of things.
This is only one of many. We could also
benefit from looking
at discussing as a form of work, as a
form of doing something. I would think that those who feel only
coding is doing have
long since departed from here. Unfortunately they managed
to paint the talkers negatively before leaving, in effect
closing off discussions. Having written an operating system in
1964 as well as participating in the writing of several others
around the same time, I got out of that mode with the
introduction of OS/360. I watched evolve from three
compatible forms (single task, multi-fixed tasks, and
multi-variable tasks) to its current MVS form. There has yet
to occur another operating system including all of the UNIX
variants altogether which can even come close.
Systems like that do not come cheap. Even a clone of OS/2
will not come cheap. How you design it will determine how
easily you can maintain it. It's obvious over its lifecycle that
its maintenance effort will exceed its development effort by
orders of magnitude. It was the cost maintaining OS/2 relative
to the value received from it that caused IBM to withdraw.
Open source has no different a people resource requirement
than closed source, except that with fulltime people you can
get the same amount of work done is less time than with
parttime. If you can't keep up with fulltime, if you can't
afford the people to keep it up, then you can expect even to
do less well with parttime. Having greater numbers of
parttime people makes the situation relative to project
management worse instead of better.
So until you can, one, afford to develop an OS/2 clone, and,
two, maintain it, then you save yourself considerable by
discussing it at length until you can do both. Otherwise you
are not simply treading water, but have the tide washing you
farther away.
REACTOS is written in C. That's not so much a crime as it is a
severe restriction. The crime lies in things like the emx
programming environment. C can't overcome those
deficiencies. As long as C is bound by unnecessary compiler
restrictions, then the open source community will have to
commit more people resources than closed source. This is one
instance where more is not better.
It is ridiculous, for example, in the age of fourth generation
languages to have to rely on beta testers for finding logical
errors in software. Then again C is not a fourth generation
language. Nor is C++. Nor JAVA. Nor Python. Nor Perl. Why
talk then of predicate logic in logic programming where the
software as part of its exhaustive true/false proof generates
the test data for every possible circumstance for all paths
within an application?
Instead of taking C on a horizontal path to another third
generation variant like C++ or JAVA, I'd rather see it go
vertically into a fourth generation form in which I can do less
and the software more. That's something we can do which
then make what we really want to do, sustain an OS/2 clone,
possible.
If you have any doubts, then follow the progress or lack of it
with REACTOS. "
I also posted it at http://groups.yahoo.com/group/osfree but have yet to receive a reply.
OS/2 Warp-ed to the very end.