SPAM: Score 3.7: Re: Instead of freebsd. com, why not...

Sander Vesik sander.vesik at gmail.com
Thu Feb 17 03:08:13 PST 2005


On Wed, 16 Feb 2005 04:05:29 +0100, Anthony Atkielski
<atkielski.anthony at wanadoo.fr> wrote:
> Johnson David writes:
> > What I am arguing is that those differences are going away. We no longer
> > live in the 1980s (when I started using BSD UNIX). While today's server may
> > have failsafe hotswap hardware, they are not inherently more powerful or
> > speedier than the clients they serve.
> 
> This is unimportant from a software standpoint.
> 

No, its not. 

> > This is what I meant by "convergence". These differences are not as
> > black and white to younger generations by virtue of the fact that they
> > are no longer black and white. While MVS may be unsuitable for the
> > desktop and OSX unsuitable for the mainframe, there's a huge middle
> > ground that includes FreeBSD.
> 
> All UNIX systems are pretty clearly servers, although some are pressed
> into desktop roles, just as all Windows systems are pretty clearly
> desktops, although some are pressed into server roles.
> 

Clear to clarify? Specificly, care to compare an average 1980s unix to
WinNT, especially WinNT 3.5 and the modern enterprise versions and
point out which part of which is more suitable for what?

> > Why is there a conflict?
> 
> Because the requirements of a desktop directly contradict those of a
> server.  And this isn't going to go away.
> 

In such a case, *NAME* those requirements. 

> > Why can't desktops be secure?
> 
> Security conflicts directly with the needs of most desktop users
> (user-friendliness, broad compatibility, broad support of network-based
> features, etc.).
> 

Security does not conflict with the needs and in reverse, is in many
cases mandated by corporate security policies anyways. Starting right
off with "no USB storage devices allowed".

> > Why can't servers have usability?
> 
> Don't they?
> 
> > Speaking of GUIs, the mere existance of /usr/ports/x11/xorg-6.8.1 does
> > not affect the performance or reliability of a FreeBSD server. Neither
> > does actually installing it.
> 
> You're half right.  Installing it destabilizes the server, and requires
> making compromises on security.
> 

Installing itself does not destabilize a server as it doesn't imply
running an X server while teh sever is in production.

> > The fact that I run KDE on my FreeBSD desktop in no way affects the
> > performance or reliability of your server.
> 
> True, but it affects the performance and reliability of your FreeBSD
> machine.

No it doesn't. 

> 
> > You can have both. Heck, you already DO have both!
> 
> It depends on how well you want to do something.
> 
> > The current desktops for X11 might not be perfect, and they might lack
> > somewhat in the usability departments, but that is no argument to
> > eliminate them.
> 
> Nobody has suggested that they be eliminated.
> 
> > The problems that FreeBSD on the desktop faces are not about the scheduler,
> > or memory management, or resources, or anything like that. Instead it's
> > mostly about getting new drivers for consumer hardware, and a little bit
> > about smoothing out the installation and configuration workflow (which would
> > benefit both sides).
> 
> Servers don't need drivers for consumer hardware, and adding lots of

It depends on what the server is made up of. 

> drivers destabilizes the OS unless they are uninstalled by default.
> Server sysadmins don't need to smooth out installation and
> configuration, since they already know what they are doing.  Smoothing
> things out means taking a lot for granted and doing it behind the user's
> back, which may be acceptable for desktops, but is often dangerous for
> servers.
> 

Than you for demonstarting yet again that you are completely out of your mind.

> --
> Anthony
>


More information about the freebsd-advocacy mailing list