[nycbug-talk] another. . .?

[Jay Savage]

On 4/26/05, Bob Ippolito <bob at redivi.com> wrote:
> 
> On Apr 26, 2005, at 1:10 AM, George R. wrote:
>
> Technically?  Not too many things... There are some bootstrapping
> issues to an installer.. you don't get to rely on a large and nicely
> configured environment being ready, you might have to fit on a single
> floppy disk, or be able to load yourself from a series of floppy disks,
> etc.
> 
> The social/psychological issues are much more of a factor.  It's
> especially hard to find somebody who wants to write an installer.
> Anyone capable of writing one doesn't need it themselves, so the
> motivation has to come from somewhere else.  Once you get that far,
> testing installers in general is a real pain in the ass, and OS
> installers are just about the worst (in user space, anyway).
> 
> > I wouldn't doubt the intentions of the creators of PC-BSD, but are we
> > really out to conquer the desktops of those who threw out their
> > typewriters last year?  With no sarcasm, I think that OS X has
> > reasonably done that, in that it's probably the most intuitive desktop
> > system around.
> >
> > And if the BSDs are too difficult for certain sysadmins, they should
> > probably find another line of work, preferably something without any
> > remotely technical content.
> 
> Seriously..  The FreeBSD installation process is cake.  It's basically
> self-documenting.  I think it's great, I wouldn't change a thing other
> than making it easy to do via PXE.
> 
> -bob
> 

I think that's a pretty Rosy view.  For most people, "install" covers
what we think of as post install, as well.  That's mostly, I suppose,
because most people think of the system as including large parts of
the userland as well.  We call a *BSD (or any unixish OS) "installed"
once we have a prompt and a network.  And then the tinkering and
configuration begins. But most people looking for a desktop machine,
don't consider the installation complete until they have a fully
functioning X environment with KDE/Gnome and common apps to perform
common tasks up and running.  Getting to that point can sometimes be a
long and laborious process.  And then once you're there, let's face
it, installing X software can be a pain.  The "installation" itself
isn't usually too bad, but because there's no default windom manager,
there's no default process for makefiles to use to add entires to
menus, etc.  so even after I've successfully "installed" a program,
I've still got several more minutes--maybe longer--of configuring the
program to work with my wm, or configuring my wm to work with the
program, or possibly both.

On Windows, whatever its many faults, adding icons to the start menu
and the desktop are options in the installer.  OS X, same thing, and
better still, programs tend to install into one directory, not one of
+- six.

Honestly, I would have a difficult time recommending any current *BSD
as a production desktop environment in a corporate or even academic
environment that wasn't greared toward some kind of computing or
technology to begin with.  the BSDs aren't, at the moment, systems for
people who need to word process, use spreadsheets, and check e-mail,
and not much else, out of the box.  Heck, at the moment, there isn't
even a reliable cross-platform integrated office suite--although the
open office ports keep getting better and are rapidly catching up with
the Linux and Windows OO.o versions.  At least on FreeBSD.  OpenBSD
doesn't even have a workable OO.o port yet.

On the other hand, I would unhesitatingly recommend Ubuntu, the latest
Kabuntu, or the new Libranet to anyone looking to replace either to
salvage old hardware or for ethical reasons, because they're not just
an OSes with x servers thrown on top, they're built form the ground up
to support a graphical environment.  And yes, part of that is an
integrated user experience from the install froward.  I would also put
SuSE on the list with the Novell desktop, but I'm frankly a little
disturbed by where that project seems to be headed.

Although I find the title of the project unfortunate, I think pcBSD
could be a huge step in the right direction for BSD.

Best,

Jay Savage