Re: [tlug] disk configuration: drive mounting

["Stephen J. Turnbull" <stephen@example.com>]

>>>>> "Marlo" == Marlo Rocci <comslave@example.com> writes:

    Marlo> Most end users want point and click solutions to
    Marlo> configuration issues,

I disagree.  They want "install and forget" solutions.  For basic
networking and so on, most Linux distros are quite close to that.  You
don't even have to remember how to reboot ;-).  Accept all the
defaults, and you'll probably have a pretty functional system when
you're done---as long as the environment is properly set up (eg, DHCP
servers and so on).  With modern concepts like WebDAV, whiteboards,
and wikis now available, it's not clear that it makes much sense to
expose "shares" to ordinary users anymore, period.  There are better
ways to publish most resources.

Obviously, you're still in the Windows world, and can't afford to
think that radically this week.  But maybe we open source developers
should.  We're never going to catch up with Microsoft if we let them
define the goals---they'll always be one definition ahead.

    Marlo>  and at this stage of the development of linux, I think
    Marlo>  they're entitled.

I bet that the great majority of Linux users, end, start, or middle,
were reasonably satisfied with the tried and true Unix text config
files (fstab, termcap, printcap) until a very short time ago---I don't
know where you come up with this "at this stage they're entitled".
Jim Breen made it plain that fstab is no big deal, even though he
doesn't want to be an admin.[1]  I think most old timers do Linux
precisely because we _want_ to be admins.  We don't want to be limited
to what Microsoft, and even more limiting, the MCSEs who run our
systems, are willing to provide.  If that means editing /etc/fstab, so
be it.[2]

The Linux community doesn't yet know what the New (ie, no sys admin
background) Users want, except Windows emulation---and I think it
would be wrong to offer that, when you can get the real thing easily
and cheaply.  We'd make even worse design and implementation mistakes
than the ones Microsoft made.  Windows interoperability is another
matter.  It would be nice if we could read Word and Excel files, and
browse the network neighborhood, etc.  But Microsoft deliberately
makes that hard, even for its own users (unless they buy regular
upgrades)!

OTOH, if you use the standard Unix model for "shares" (NFS, AFS, Coda
for file systems, lpd for printers, etc) you could offer "install and
forget" pretty easily, I would think.  Unix admins have been doing
this on a site-by-site basis for decades.  You add very standardized
stanzas to fstab and printcap and config files like that.  This could
easily be automated by a trivial script, or by a network copy using
ftp or scp, or a network boot with automounted /etc.

And note: no problem at all for the users.  Eg, browsing for "user
shares" is transparent: "ls /home | less".  Just mount all the user
homes via NFS.  All the GNOME/KDE graphical tools will do this "right"
too.

But it sounds to me like you want to assume that Linux is a guest in a
Windows world.  But why would Linux admins and developers write a tool
from that point of view?  Not only are they biased in favor of
assuming *nix, it's a lot easier to write tools for *nix.  Not least
because most *nix subsystems (as a matter of design) can be configured
fairly independently of others.  This makes it easier for the
experienced admin with a fairly unique task to whip up a solution,
while making things harder for a less experienced or overworked admin
to fit the *nix box into a bondage & discipline-oriented site.  And
_much_ harder for a "auto-configurator" to do so.

    Marlo> What I am seeking is to what extent has linux progressed in
    Marlo> that direction.

Not very far.  Most of the projects I'm aware of are interested in
interoperability, not in making Linux a "native emulation" of Windows.
Samba has some neat stuff and there are clients to do autoconfiguration
(ISTR) and neighborhood browsing (not the Lin-neighborhood somebody
else mentioned, that's another tool),[3] but I doubt you'll find good
support in generic Linux config tools.[4]

    Marlo> The impression I am getting from the users

Your Linux users?  If that's what you mean, maybe you should listen to
them and let them do most of the work---probably at least some of them
enjoy it.  Look over their shoulders a bit to make sure that what they
come up with is compatible with "policy", whatever that means in your
installation.

    Marlo> is that I shouldn't even be asking the question, and I
    Marlo> should be happy with editing text files.

Probably true.  Windows networking works as well as it does for one
reason:  Microsoft's ability to enforce policy on Windows networks.
There's one "share" model, one identification model, one
authentication model, etc, etc.  *nix networking is inherently
heterogeneous, and mixing in Microsoft hosts only makes it worse,
since Microsoft generally made different choices in most dimensions.

If you know what you want, then it's probably best for _you_ to write
a script to produce that result.  The fact is that it has taken
Microsoft more than a decade to produce tools that do half the job
that "su; cd /etc; for i in *; do pico $i; done" can do, and it takes
about the same amount of time.  vi would make the comparison totally
unfair, of course.  :-)


Footnotes: 
[1]  He says out of the box fstabs are never what he wants.  Well,
I've never seen a Windows installation that didn't need a fair amount
of tweaking before it worked, let alone started to work the way I
wanted it to.  Things that are easy to do with Linux systems require a
reboot or even reinstall on Windows.  These systems are just aimed at
different kinds of users.

[2]  Oh, and did you ever try to admin a mixed Novell/Lanman network?
It's enough to make one fall madly in love with hand-editing /etc/fstab
---or even /usr/src/linux/.config---it is.

[3]  Despite the fact that the raison d'etre of Samba is really to
provide SMB servers that don't "fall down and can't get up", the
clients work well, too.

[4]  That would be like asking Microsoft tools to stop exporting
Windows-specific proprietary document formats on the Internet, and
start producing conformant MIME, XML, Java, etc, etc.


-- 
Institute of Policy and Planning Sciences     http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba                    Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
              Don't ask how you can "do" free software business;
              ask what your business can "do for" free software.