NetNews Usenet Archive 1992 #20/NN_1992_20.iso/spool/comp/unix/xenix/sco/2907

Text File | 1992-09-12 | 5.6 KB | 108 lines

Newsgroups: comp.unix.xenix.sco
Path: sparky!uunet!haven.umd.edu!decuac!pa.dec.com!dynamix![e-mail]
From: [e-mail] (David L Jarvis)
Subject: Re: Xenix considered harmful (was Re: SCO support - a success story)
Organization:  SOFTWARE / DYNAMIX
Message-ID: <[e-mail]>
Date: Sat, 12 Sep 92 13:28:35 EDT
X-Received: by usenet.pa.dec.com; id AA25239; Sat, 12 Sep 92 10:38:01 -0700
X-Received: by inet-gw-2.pa.dec.com; id AA27725; Sat, 12 Sep 92 10:38:00 -0700
X-Received: from dynamix.UUCP by uu3.psi.com (5.65b/4.0.071791-PSI/PSINet)id AA27983; Sat, 12 Sep 92 13:33:50 -0400
X-Received: by dynamix.com (smail2.5c)id AA27150; 12 Sep 92 13:28:35 EDT (Sat)
X-To: comp.unix.xenix.sco.usenet (comp.unix.xenix.sco)
X-In-Reply-To: <[e-mail]>; from "Marc Unangst" at Sep 6, 92 8:32 pm
X-Mailer: ELM [version 2.3 PL11]
Lines: 91

> Well, we're picking nits here, but I can't resist.  During most of my
> experience with Xenix/286, you'd get a machine lockup more frequently
> than you'd get a segvio.  Segvios were more frequently caused by
> a large-model program that tried to assume pointers and ints were the
> same size.

So ... first you question me as to why I feel that Xenix on a 286 isn't a
stable computing platform, then when I give you my answer, you "pick nits"
and tell me *your* reason why it's not stable.  Initially you were arguing
that it was, now you contradict my reason with one of your own and tell us
that it really isn't.  Then again, below, you contradict yourself again
concerning your initial argument that Xenix has no reason to exist.

As to your "nit picking" above ... "during most of my experience with
Xenix/286..." doesn't mean very much coming from someone
who's *obviously* had *VERY* little experience with Xenix at all,
much less a very old version of it ...
I was in Xenix tech support for Tandy when 286 Xenix first came out ...
(about 7 years ago)  I saw *huge* amounts of Xenix/286 boxes go out the
door, and come right back in again, and sit on my desk ... so, in *my*
experience, I saw more "segvio"'s than lockups, but we did see those too!
(segmentation violations may have been more prevalent due to the fact that
most went out with the same applications programs)  But to argue that the
286 & Xenix was unstable for one reason over another is ridiculous ... you
questioned me as to why I felt that way, and here is my answer ...

> Well, we haven't switched them to Unix because they don't have the
> money and because there was really no need.  However, [.......]

here again you contradict yourself by admitting that there are indeed
*real* clients out there who don't need Unix in all it's glory ...

> I'd hardly call Xenix "refined".  "Old", maybe, but old does not
> mature make.

I don't think what YOU would call Xenix matters much to most ppl reading
this group ... you've already proven you don't know squat about Xenix and
it's proper administration and usage in the business world ...

> Uh, there are lots of things you can do with inittab that you can't do
> with /etc/ttys.  Like spawn something that's not called "/etc/getty".
> There are other things I could go into, but they have more to do with
> the SysV-style rc scripts vs. the V7-style rc scripts.

I said *normal* circumstances ... how many of your average business
installations of *nix involve running something other than
/etc/getty on a port?

> You obviously haven't ever tried to add more than one or two
> third-party drivers to a Xenix system.  I have.  And then I had to
> clean up the link_xenix script with an editor because the drivers had
> each stomped on each other.

Sure I haven't ... you know this, because, well, you just do ...
so why don't you tell us all then, oh-wise-and-experienced one, just what
drivers you claim don't install right on Xenix?  Remember, different
versions of drivers are required for different versions of Xenix!
I have *NOT ONCE* seen a driver developed for the proper version of Xenix
that had trouble installing.

> Sure, you *can* get domain-based e-mail running on a Xenix system, but
> it's an unbelievable amount of trouble to rip out the Xenix e-mail
> system and replace it with your own.

I can't believe you would make such ridiculous claims in the company of ppl
who are *obviously* lightyears more experienced with Xenix than you.
Again, why don't you tell us exactly what you've had to "rip out" of Xenix
to get domain-based email running???  I can do smail 2.5 in 10 minutes and
not even break sweat, much less have "unbelievable amounts of trouble
ripping out Xenix email" ...
Or was your point more along the lines of, if we make our clients buy SCO
Unix instead of Xenix they'll never have to use anything other than SCO
programs installed with the system???  (equally ridiculous)

> it interoperates well with their other Unix machines.  Xenix *may* be
> appropriate if interoperability is not and never will be a concern,
> but I don't believe *anyone* can say that in this day and age.

Right, everybody knows that Xenix can't do UUCP, TCP/IP, or any of that
other "interoperability" stuff ... (theres another buzzword)
Also, many of our clients dont have the need to interconnect different
machines (if they interconnect at all, its on a minor scale, ie: dos-unix,
or it's with similiar machines, ie: sco unix - sco xenix) ...
many dont even have more than one!  (at least the ones we've
put on Xenix)


#----------------------------------------------------------------------#
#          David L. Jarvis                 SOFTWARE  /                 #
#          [e-mail]                       /  DYNAMIX          #
#----------------------------------------------------------------------#