NetNews Usenet Archive 1992 #27/NN_1992_27.iso/spool/comp/sys/sun/admin/8505

Text File | 1992-11-20 | 3.9 KB | 80 lines

Newsgroups: comp.sys.sun.admin
Path: sparky!uunet!utcsri!geac!censor!comspec!noweh.com!georgn
From: [e-mail] (Georg S. Nikodym)
Subject: Re: avoiding Solaris
In-Reply-To: [e-mail]'s message of 16 Nov 1992 11:01:59 +1100
Message-ID: <[e-mail]>
Sender: [e-mail] (Georg S. Nikodym)
Organization: Noweh Software, Mississauga, CANADA
References: <[e-mail]> <[e-mail]>
    <[e-mail]>
Date: Fri, 20 Nov 1992 05:02:54 GMT
Lines: 66

In article <[e-mail]> [e-mail] (Charles Meo) writes:

   Had Berkely given more thought to a sensible system administration interface
   (and I'm not saying SVR4 has one, mind you!) this situation need never have
   happened, since the the substance of all the criticisms of old-style UNIX
   is that the machines are too damned hard to run.

You ask a great deal of the 4 or 5 people that formed the core of the
Berkely effort (of course, I'm ignoring the thousands of contributions
from other places).  My point is that Berkely was always largely a
research effort and the people using it are generally technical enough
to accept any shortcomings in the sys-admin side of things (and those
that weren't have become so).

   Unfortunately, IMHO the way SVR4 is being marketed you'd think that people
   no longer need anyone in their datacentre who knows what they are doing--
   anyone who thinks running a 600-user supermini will be as easy as running a PC
   in the brave new world is in for a nasty shock.


Tell me about it...  BUT, that's pretty much the attitude that MIS types
have held when the PC was born.  And the expression of that attitude that
has driven the user community to seek alternatives to the datacentre.
Remember that we computer people work for/with people not technology.

   There are too many black boxes in SVR4--people who do know what they are
   doing cannot enforce intelligent configuration decisions even if they know what
   they are. This seems stupid to me. Lots of the config parameters are
   undocumented, many more are 'Oh, you don't mess with that, ...  will break'
   It boils down to about six things you can usefully mess around with.
   Nobody is going to tell me that an OS designed to run from PCs to mainframes
   is always going to use the right strategies, and can do a better job than
   an experienced admin who knows his/her hardware and workload. And yet it's
   pretty much a case of, if it doesn't do what you want out of the shrinkwrap,
   tough.

You raise a number of good points.  However, the goal of making computers
easier to use, software easier to install, etc is still a noble one and those
working towards it need not be cynically dismissed as wasting time.  Maybe
you'd like to boot your machines by entering the boot code via a switch panel
again?

   Also, I would have to say that I have found ATT to be somewhat unresponsive,
   shall we say, to bug reports and fixes emanating from organisations who have
   had to deal with (sometimes very imaginatively broken) earlier releases
   of System V. There is no indication that this will change.

This is a valid concern and one of the many reasons folks (and corporations
large enough to be able to afford a differing opinion) have a problem with
the way AT&T (now USL) has managed the emergence of UNIX as one of the
major operating systems in common use.

   Anyone from Bell Labs care to respond? In fact, I am willing to be persuaded
   on this issue. Anyone at all been able to do anything outstanding with SVR4?

I suspect the people in Bell Labs don't give a rat's ass about UNIX
problems for two reasons:

    1)  They play with Plan 9.
    2)  The Labs were the birthplace of UNIX,
        but they are not the current owners.
--
Georg S. Nikodym  -  [phone] / 720-4729
Noweh Software - Mississauga, Ontario, CANADA
UUCP:    {comspec.com, lsuc.on.ca, uunet.ca}!noweh!georgn
RFC822:    [e-mail]