[haiku-development] Re: Haiku future IMH point of view
- From: "Michael Phipps" <michael.phipps@xxxxxxxxxxx>
- To: haiku-development@xxxxxxxxxxxxx
- Date: Wed, 10 Sep 2008 07:15:02 -0400
Hi Daniel!
This one should be an FAQ. :-)
Binary backward compatibility is a good idea because it gives a reference
implementation of an API. It lets developers check their work against a "known
good" solution. It gives Haiku developers a large (relatively) base of software
to test their OS code with. It ensures that the API is correctly implemented.
Last time I asked the developers how much time and effort backward
compatibility took, their answer was that it was minimal, except for the work
on gcc 2.95.3. In other words, the system would not be further along by much if
they dropped backward compatibility.
Your suggestions for platforms to port to are interesting. It is assumed that
ANY port of the OS would not have any binary compatibility. Certainly that is
true for the PPC port that a few people are interested in. Keeping x86-32
binary compatibility does not IN ANY WAY impact ports to other platforms. The
additional ability to test and the lack of "discussion" about how things
"should be" are such a powerful development and testing incentive that I would
even argue that retaining binary compatibility HELPS ports.
Finally, it was never part of the plan to retain binary compatibility forever.
It was only ever intended to be PROMISED for R1. It may happen for later
versions depending on dev interest and on the pain that is caused by switching
(i.e. how many apps there are without source that would be lost by dropping
it).
All, of course, IMHO and without any ability to speak for anyone other than
myself.
Michael
-----Original Message-----
From: "Daniel Oliveira Costa Lemos" [xspager@xxxxxxxxx]
Date: 09/09/2008 20:15
To: haiku-development@xxxxxxxxxxxxx
Subject: Re: [haiku-development] Haiku future IMH point of view
Hello everybody!
Please, someone correct me if i commit any mistake, but the gramatical ones ;)
Thinking about what would be the differential of Haiku, if you take of
the BeOS binary compatibility. I think the hype about BeOS, at last in
my opinion, of whom never use BeOS, was, the support for SMP and easy
use o threads, be programed in a Object Oriented language (C++), and
it's clean design and legacy free.
IMHO I think Haiku should do what BeOS did in the past, supporting the
newer hardware platform, and the last technology in hardware for
desktop. So, I think would be an advantage to Haiku to support
AMD64/EM64 in the full 64bits mode, and possible a machine with a Cell
chip (probably PS3) too. Maybe and possible not for now, but some
point in the future, we would be able to extract maximum performance
from this architectures using Haiku due it's design that comes from
BeOS. If you could, imagine BeOS running on a Cell CPU on Sony's PS3.
All this SEPs ... :D
Sorry if i don't make me understand or if I'm wrong, but my point is
about the future of Haiku, that it will do what Linux can't do, or be.
Be a high quality and integrated, open source graphical desktop
operating system.
Other related posts:
