Re: [PLUG] ESR: Last phase of the desktop wars

[Rich Freeman via plug <plug@lists.phillylinux.org>]

On Mon, Oct 12, 2020 at 12:13 PM JP Vossen via plug
<plug@lists.phillylinux.org> wrote:
>
> Listing new Slashdot and original links:
>
> 2020-09-25: ESR's "Last phase of the desktop wars? http://esr.ibiblio.org/?p=8764
> * https://linux.slashdot.org/story/20/09/27/193250/eric-s-raymond-is-microsoft-switching-to-a-linux-kernel-that-emulates-windows
> * https://linux.slashdot.org/story/20/10/11/0225229/what-if-they-replaced-windows-with-microsoft-linux
>
> The GUI is not hard, you can already theme various Linux to look close enough.  These things might be harder:
> * "/" vs "\"
> * LF vs CRLF (most editors handle this now, but...)
> * Where did "C:" and "D:" and so on go?  (Good riddance!)
> * Right now, ALL the apps, esp. games, but see TFAs.

I think a lot of things are harder:

* Stable kernel module ABI - they'd probably have to fork the kernel
just for this, and at that point why bother?
* Full support for non-GPL kernel modules.  The Linux devs are trying
to make it harder to use them.  I can't see MS following that path.
At best they'll wrap one giant GPL condom around the whole thing, and
then we can find out if APIs really are copyrightable like Oracle
claims (I really see no practical difference between the claim that
you can copyright an API and the claim that dynamically linking
creates a derivative work in the linked binary and not just the
ephemeral runtime-linked memory image).
* Ability to update your video drivers without restarting any
graphical apps.  Already near-seamless in Windows, and Linux barely
can handle small kernel patches without a reboot - it is too
monolithic.

Don't get me wrong, linux has a lot of benefits, but it isn't like
there aren't things Windows can do that Linux can't.

Legacy support is another potential area with issues.  The kernel
itself is fine as long as you add the stable driver ABI - the
userspace interface is already very stable.  Stuff like X11 is ok
because it is basically legacy code at this point.  However, MS is
going to want any OS-like capability in their distro to have a stable
ABI for 10 years after obsolescence.  Ie, WinLinux 2020 comes out, and
binaries have to run without modification until 10 years after
WinLinux 2028 comes out - ie 18 years after launch.  That's what
they've been doing for the most part until now, and it is why
everybody uses them in the Enterprise.  RHEL/CentOS are popular on
servers precisely because they're the closest you can get to this on
Linux.

I feel like by the time MS forks Linux enough to make it into what
they would want it to be, all the Linux purists wouldn't want to touch
it with a 10' pole.  Just look at Android or ChromeOS - they're "100%
Linux" and have little to do with the typical Ubuntu experience.
Plus, all those drivers people end up creating for WinLinux would end
up not working on regular Linux because Linus will have changed the
kernel ABIs 47 times by the time release 1 comes out, so that will
just tick everybody off.

And with the amount of forking they'd end up doing, why start with the
Linux kernel in particular?  Why not pick something that is more
modular, like just about every modern alternative?

-- 
Rich


-- 
Rich
___________________________________________________________________________
Philadelphia Linux Users Group         --        http://www.phillylinux.org
Announcements - http://lists.phillylinux.org/mailman/listinfo/plug-announce
General Discussion  --   http://lists.phillylinux.org/mailman/listinfo/plug