On Mon, Apr 18, 2016 at 3:51 PM, Jim Oberholtzer
<midrangel@xxxxxxxxxxxxxxxxx> wrote:
So the better question is "what's an open system"? IBM i certainly falls
into the "open" category for my uses.
I guess that might be a better phrasing, but that *WAS* my question.
I was trying to find out what Rob meant by "open system" by learning
how IBM i meets the definition, and how DOS/Windows does not. But
yeah, a straight-up definition would probably be even better.
Based on the responses that I got, it seems that the openness of a
system is not a you-have-it-or-you-don't quality, but rather a matter
of degree. Fine, I think that's entirely reasonable.
So the various ways in which a system can be considered open included
"having lots of published (and preferably OSF-sanctioned) APIs" and
"being able to build and run open-source software". I think when you
really boil those two down, what you get is that a system is "open" to
the extent that it adheres to POSIX. This is more-or-less in line with
the Wikipedia entry on open systems:
https://en.wikipedia.org/wiki/Open_system_%28computing%29
From a less technical and more practical standpoint, I would say
DOS/Windows is not *particularly* closed. At least not all that much
more closed than IBM i. IBM i is not too open or POSIXy without PASE,
and you can get MSYS or Cygwin for PCs (for free) which will serve
much the same purpose as PASE does for i. And even without POSIX
layers, PCs can run plenty of open-source software.
I would also say that "proprietary" is not a good antonym for "open",
at least when trying to make the case that IBM i is open.
John Y.
midrange.com
