18 Jan
2013
18 Jan
'13
2:34 p.m.
Before you put too much hope on it, ReactOS needs a Memory Managing
Unit (basically, virtual memory support), which is present only on
ARMv7. Raspberry Pi has an ARM11 CPU, which only implements
ARMv6. (yeah, don't be confused with all those numberings...)
Le Thu, 10 Jan 2013 21:41:10 -0600,
"J. C. Jones" <jaibuduvin(a)gmail.com> a écrit :
Hi All,
Before I reach the point where I am thinking the same way as everyone
else, which will inevitably happen (The Borg - resistance is futile),
I wanted to get some thoughts out:
A question that many have asked is, "Why not port ReactOS to ARM?"
The answer is usually something like, "We cannot afford the resources
to port to ARM."
I think on the contrary, the opposite might be true:
1. There is a hoard of developers over on the Raspberry Pi
site right now who would enjoy seeing ReactOS on the RaspPi.
2. There are device manufacturers who would like to free
themselves from the Apple/Google/Microsoft triumvirate and
iOS/Android/Windows Phone 8 lock-in. Samsung recently announced its
intent to explore other operating systems.
http://news.cnet.com/8301-1035_3-57455054-94/would-samsung-ever-leave-androi
d-new-ceo-drops-hints/
3. The operating systems are not exactly easy to develop for. I
have read credible articles that Android is a mess from a development
perspective. ReactOS would be the operating system of choice for
straightforward development.
4. There are embedded systems companies who struggled in vain
to get Windows CE to behave like "Big Windows", but were unsuccessful
because restrictions in the Windows CE kernel. Many of them switch to
Linux, but quite a few still use some form of embedded Windows, and
would welcome an open-source Windows-like OS.
5. The United States Military has very large base of software
that they would like to put on lower-power systems (ARM) that is
written for Windows/i386. They are currently trying to port this
software to Linux, with varying degrees of success, not because they
like Linux, but because they need as much of the software to be open
as possible. They would be particularly attracted by the open-source
nature of ReactOS, because the USA national security vetting process
requires that certain classes of software be reviewed, line by line,
by a certain US security agency. The singular, totally exposed nature
of ReactOS makes it attractive in this regard.
6. ReactOS.ORG would likely receive real money from device
manufacturers. Even a few dollars per-device would add-up very
quickly.
7. There is NO mobile platform right now, among the Big Three,
that allows true, native, C/C++ development. Each of these platform
plays a game where the native code is invoked by some shell, even in
the case of Windows Phone 8, despite Microsoft's claim that Windows
Phone 8 supports native development. [It depends on what your
interpretation of native development is.]
ARM device manufacturers are all stuck in the same boat. Most of these
companies are actually not very good at OS design. Think about it:
Nokia was a multi-billion-$US company that was using an operating
system (Symbian) that was so broken and toxic to innovation that it
almost drowned their company. What did they do to fix this problem?
They adopted a closed OS from Microsoft. Manufacturers, actually, do
not like having closed software. It eliminates their opportunity for
differentiation. If ReactOS were made to run on a single
manufacturer's device, the other manufacturers would become nervous,
and insist on having the same access as does their competitor. There
is nothing wrong with making these manufacturers pay a small fee to
support the ReactOS Foundation, and they would gladly pay it, if we
developed killer applications for their devices.
Of course, because most of ReactOS, in theory, should be portable,
software working on ARM is software working on x86_32/x86_64. I would
also like to mention here that there are a lot of developers who
would much rather have a stable kernel, and a paucity of user-mode
applications, versus an unstable kernel, and a plethora of user-mode
applications. User mode applications will be created by the hoard,
*if* the kernel is stable. If the kernel is not stable, the incentive
to do anything else is greatly reduced.
This is the opportunity I see. My biological clock is telling me that
2013 is the year to pursue this effort. The market is waiting. But an
effort like porting to ARM should not be done haphazardly or
opportunistically. It should be done with deliberation and intent.
Just my opinion.
-John