From owner-misc+M40684@openbsd.org Sat Aug  9 03:56:40 2003

	From: Chris <racerx@makeworld.com>
	Subject: Think outside of the box (Gawd, I hate that term)
	Sender: owner-misc@openbsd.org
	To: freebsd-questions@FreeBSD.ORG
	Cc: netbsd-help@netbsd.org, misc@openbsd.org

		As great as the *BSD's are - why are we not thinking ahead of the curve? 
	Outside of the box? What do I mean? Well - why not develope our OS's to run 
	on an arch. that isn't out yet - MAKE that new arch. just like Gates did so 
	many years ago. He defined the x86 market - why can't we do the same?

Alright, I see some problems with that.  It would be a great idea, but
for some very important considerations.  I'm on the OpenBSD side, so I
would not be able to talk about the other BSD's, but here's what I
think.  

One, the goals and motives of the project are not for building a
new box, new architecture, or new machine, they are purely software
oriented, and they try to support many architectures.  They have done
this fairly well, but not nearly well enough that they can all of a
sudden branch out into other territory without proper funding, and
without a reason.  

Two, the whole object of that would be to make money,
and at the moment, the PC market is rather saturated, making it
difficult for new-comers to enter into the market successfully; while it
would be feasible, I doubt that it would be an easy task, and the BSD's
don't have the funding for it, when it could be better spent elsewhere.

Three, Gates isn't that way, he didn't design everything, he took,
borrowed, and dealt, then tyrannized.  That's what we call a dictator,
using whatever means necessary to get to the topic [he would have done
just as well with communism or socialism] and then staying there.  I
don't think we wish to mimic him.

Four, the amount of money, time and skill it would take to develop such
a new architecture is enormous, this cost is so great, that almost no
company has done this in years.  IBM has stuck with PowerPC and x86 for
years, SGI has been with MIPS for as long, relatively; even BeOS or Sun
have not made new architectures.  These architectures were developed
years ago, and then fine tuned.  It was a huge effort to make them, and
then to fine tune them to the point that they are now.  No company that
I know of now is actually developing a new architecture, just
modifications and upgrades of the already somewhat good architectures.

Five, while I would love to see a new architecture come in, if it was
better, cleaner and faster, I do not think it wise to expect a software
company(ies) to be able to do this.  If you want to do this, use the
grand capitalist approach, work on gaining some funds, then begin the
development stage, then you can market it.  It would be much better if
this was a project wholly individual of the other systems, as it would
make development more centralized, and better organized, more focused
and capable.  Having the BSD's do it is like telling a woodworker to
start painting with oils, or vice versa; and expect a good result.  The
only people I know of that can actually do this, is Apple.  But they
have an amazing bunch of team out there, and have massive amounts of
funds pouring into R&D; not to mention, they started really early, and
their company is still a hard company to run and work.

To sum it up, it's not wise to get the BSD's to do a project that would
be better done as a separate entity, and then to port BSD over by
donating the proper funds to the BSD's.

Aaron Hsu
http://www.aaronhsu.com