From: Conrad Sanders <sandc...@ix.netcom.com>
Subject: SCO Open Server VS Linux
Date: 1996/11/29
Message-ID: <329F11FB.27E1@ix.netcom.com>#1/1
X-Deja-AN: 201408270
content-type: text/plain; charset=us-ascii
organization: United Mailing Inc.
x-netcom-date: Fri Nov 29 6:35:13 AM PST 1996
mime-version: 1.0
newsgroups: comp.unix.sco.misc
x-mailer: Mozilla 2.0 (Win16; I)
Has anyone out there tried Linux ? I am trying to decide if I should
install Linux or SCO Openserver. Any comments ?
--
Conrad R. Sanders
United Mailing Inc.
From: da...@hendrix.postino.com (Danny Aldham)
Subject: Re: SCO Open Server VS Linux
Date: 1996/11/29
Message-ID: <57nbt8$npn@hendrix.postino.com>#1/1
X-Deja-AN: 201441436
references: <329F11FB.27E1@ix.netcom.com>
organization: Postino DotCom
newsgroups: comp.unix.sco.misc
X-Newsreader: TIN [version 1.2 PL2]
Conrad Sanders (sandc...@ix.netcom.com) wrote:
: Has anyone out there tried Linux ? I am trying to decide if I should
: install Linux or SCO Openserver. Any comments ?
Personally I prefer FreeBSD. ;-)
Really though, why not try both (or all 3) and decide for yourself.
Free OpenServer is easier to install, IMO, and it auto installs a
basic X-window , which I found to be a PITA to configure on Linux.
After that, you will learn tons with either OS.
--
Danny Aldham www.postino.com
Need a UUCP E-Mail feed? - Ask me!
Dial-up access in British Columbia. TCP access around the World.
From: "Scot Harkins" <sc...@wolfenet.com>
Subject: Re: SCO Open Server VS Linux
Date: 1996/11/30
Message-ID: <01bbde5e$2e516940$c1629dcc@scoth.wolfe.net>#1/1
X-Deja-AN: 201487442
references: <329F11FB.27E1@ix.netcom.com> <57nbt8$npn@hendrix.postino.com>
content-type: text/plain; charset=ISO-8859-1
organization: Thurman Industries
mime-version: 1.0
newsgroups: comp.unix.sco.misc
Howdy,
Trying them all is a good idea. It is much easier to do so know with
FreeSCO. The question after that becomes:
1. Will you be storing data that is critical to the operation of your
business? If so, are you comfortable enough to take on the responsibility
of support Linux on your own, versus being able to call support at SCO
(which costs money)? I've had to rely, from time to time, on outside
assistance from my SCO vendor for questions on the several SCO systems I
manage. Not terribly often, but enough so that I'm grateful for using a
commercial Unix. I am contemplating Linux or FreeBSD for other multi-user,
non-critical functions. The systems that run our retail stores, however,
must be reliable, and that's what they are with SCO.
2. Are you prepared to deal with hardware vagaries? Commercial Unixes
feature broad hardware support, either out of the box or from hardware
vendors. That also includes technical support. Those vendors who
"release" drivers for Linux generally do not give any technical support
(there are some who do, which is cool) when you encounter a problem.
3. There are a few really good, reliable distributions of Linux that work
"out of the box", provided you stay within the range of known good
hardware. Red Hat, SlackWare, and others are good. Their installations
may also be as easy or easier than OSR5 from SCO. Linux is not a problem
waiting to happen, it's just more raw and on the edge (in terms of support,
hardware and technical). There are, however, some VAR's who actually
install and support Linux systems in commercial environments, and there are
some businesses/institutions that utilize Linux quite successfully in
"business critical" operations. It really boils down to how you feel about
your ability to handle even the hairy problems on your own.
4. Price. Commercial Unixes like SCO cost Commercial Dollar$. The addage
"you get what you pay for" applies.
5. You could make a trial run at the concept of the computerized data via
Linux. If your company comes to rely on the operation, you may then wish
to plan a migration to a commercial Unix system. They (the bosses, if any)
may say "why spend the money?" If you saved them buttloads with the system
you have, then ask them how much it would cost if the system you had went
down hard for a couple of days. Would the added reliablity and support be
worth the cost?
There's my 25 cents.
Scot
--
Scot Harkins (KA5KDU) | Systems Administrator, Thurman Industries
SCA: Ld. Scot MacFin | Bothell, WA. SCO Unix, Win, Win95, DOS 5&6.x, Mac
sc...@wolfenet.com | Native Texan. Proud and Happy Daddy. Herald at
large.
sc...@scn.org | URL <http://www.wolfenet.com/~scoth>
Danny Aldham <da...@hendrix.postino.com> wrote in article
<57nbt8$...@hendrix.postino.com>...
> X-Newsreader: TIN [version 1.2 PL2]
>
> Conrad Sanders (sandc...@ix.netcom.com) wrote:
> : Has anyone out there tried Linux ? I am trying to decide if I should
> : install Linux or SCO Openserver. Any comments ?
>
> Personally I prefer FreeBSD. ;-)
> Really though, why not try both (or all 3) and decide for yourself.
> Free OpenServer is easier to install, IMO, and it auto installs a
> basic X-window , which I found to be a PITA to configure on Linux.
> After that, you will learn tons with either OS.
>
> --
> Danny Aldham www.postino.com
> Need a UUCP E-Mail feed? - Ask me!
> Dial-up access in British Columbia. TCP access around the World.
>
From: e...@bigbird.telly.org (Evan Leibovitch)
Subject: Re: SCO Open Server VS Linux
Date: 1996/12/02
Message-ID: <E1rwI8.3ry@bigbird.telly.org>#1/1
X-Deja-AN: 201830071
references: <329F11FB.27E1@ix.netcom.com> <57os2d$h0t@comix.comix.santa-cruz.ca.us>
organization: Sound Software
newsgroups: comp.unix.sco.misc
In article <57os2d$...@comix.comix.santa-cruz.ca.us>,
Jeff Liebermann <je...@comix.santa-cruz.ca.us> wrote:
>Conrad Sanders (sandc...@ix.netcom.com) wrote:
>: Has anyone out there tried Linux ? I am trying to decide if I should
>: install Linux or SCO Openserver. Any comments ?
>: Conrad R. Sanders
>: United Mailing Inc.
>I like to reduce such questions to simple dollars.
>Let's pretend that all operating systems were absolutely free.
>You buy the server hardware and just run it. Now, what
>is the cost of one crash or mutilated database? Some
>of my commercial customers value an all day crash and
>server rebuild at 2-3 times what the server costs. One
>customer carries insurance against downtime costs.
>These customers pay for reliability and uptime because
>the very first screwup will cost much more than the
>entire operating system.
This is all fine ans well, but does not answer the question one shred.
If there is any evidence whatsoever that one of the commercially-sold
Linux distributions (Caldera or Red Hat) is less stable than SCO,
let's hear it. So far all I've heard in this realm is FUD.
There are reasons to choose SCO or Linux. Reliability is not one of
these issues. Installed according to the vendor's instructions, either
can run very well.
The main difference is that, since Linux includes source, it's easier
to muck with a Linux installation if you want. But if reliability
is an issue, who's going to go about mucking with the source?
>Linux is primarily a workstation.
I have yet to install a Linux system that's 'primarily a workstation'.
They're Web servers and POP servers and Samba servers and firewalls and
routers. That they have good workstation support (such as sound cards)
in manners that commercial Unixes may have forgotten about is but a
bonus in such installations.
>Workstations are feature intensive while servers tend to
>be reliability intensive.
>Workstations tend to have multiple operating systems while
>servers tend to have only one.
>Workstations tend to get rebooted often while servers are
>expected to stay up 24 hrs/day.
>Workstations tend to run a multitude of applications while
>servers tend to run a few big applications.
>I could go on, but methinks the distinction is evident. If
>your application and customer can tolerate the workstation
>methodology, then by all means run Linux. If your application
>and server are expect to play continuously, methinks SCO
>Unix would be a better choice.
I don't understand the leap of logic here. Nowhere do you say that Linux
is unsuitable as a server, nor that it is incapable of reliable operation.
How many of the world's web servers run on Linux? How many on SCO? Do
any of those thousands upon thousands of Linux webmasters have any
reliability problems using their systems as (primarily) servers? Is
the Web world lining up to dump their Linux systems now that FastStart
is here? Don't bet on it.
It is true that SCO has far, far more commercial applications available
for it than Linux. But for those apps running on both platforms, such as
ADABAS or WordPerfect or Wabi, is SCO demonstrably more robust?
Last week I wrote a lengthy bit in comp.unix.unixware.misc about the
distinctions between Linux and UnixWare. Much of those differences
apply to OpenServer as well, but stability is not a significant
point of difference between them.
>The next obvious question is how much better is SCO Unix.
>I would stick my neck out and say that 90% of all the Unix's
>are fundamentally the same and reasonably usable. However,
>that last 10% will make the difference between functional
>and marginal. If marginal is all you want, then Linux is
>just fine. If you expect somewhat better than that, methinks
>SCO Unix is the answer.
I usually agree with Jeff on most things, but I guess we'll have
to agree to disagree on Linux's ability to cross the leap from
marginal to functional.
There are still things I like about SCO far better than Linux,
but I'd never suggest that Linux is less "functional" in any
significant way. Compare OS5's non-Morningstar networking to
Linux's out-of-the-box stuff and tell me, with a straight face,
that this aspect of Linux is any less functional than SCO.
I still say that the day Oracle and Informix some out with native
ports for Linux, SCO has something real to worry about. The very
high-end of SCO's market will be safe, but the low and middle
ground will be very much up for grabs.
--
Evan Leibovitch, Sound Software Ltd, located in beautiful Brampton, Ontario
Supporting PC-based Unix since 1985 / Caldera & SCO authorized / 905-452-0504
Unix is user-friendly - it's just a bit more choosy about who its friends are
From: Mike Jagdis <m...@roan.co.uk>
Subject: Re: SCO Open Server VS Linux
Date: 1996/12/02
Message-ID: <E1s6tH.A6@roan.co.uk>#1/1
X-Deja-AN: 201859491
x-mail2news-path: relay-7.mail.demon.net!relay-5.mail.demon.net!
roan.demon.co.uk!toaster.roan.co.uk
references: <329F11FB.27E1@ix.netcom.com> <57nbt8$npn@hendrix.postino.com>
<01bbde5e$2e516940$c1629dcc@scoth.wolfe.net>
x-mail2news-user: n...@toaster.roan.co.uk
organization: Roan Technology Ltd.
x-comment-to: "Scot Harkins" <sc...@wolfenet.com>
newsgroups: comp.unix.sco.misc
>1. Will you be storing data that is critical to the operation of your
>business? If so, are you comfortable enough to take on the responsibility
>of support Linux on your own, versus being able to call support at SCO
>(which costs money)? I've had to rely, from time to time, on outside
>assistance from my SCO vendor for questions on the several SCO systems I
>manage.
Support usually costs extra these days and you can buy support for
SCO, Linux and even Windows.
>If you saved them buttloads with the system
>you have, then ask them how much it would cost if the system you had went
>down hard for a couple of days. Would the added reliablity and support be
>worth the cost?
If you are claiming added reliability of the OS for the price you are
going to be shelling out serious money for code fixes with a 4-8 hour
turn around. I have never actually heard of anyone even consider
such a thing (at that level a source license is attractive). If you
are just buying hardware maintenance the OS isn't particulary important.
Like you said, "You get what you pay for". All too many people just
like to pay regardless :-).
Mike
--
.----------------------------------------------------------------------.
| Mike Jagdis | Internet: mailto:m...@roan.co.uk |
| Roan Technology Ltd. | |
| 54A Peach Street, Wokingham | Telephone: +44 118 989 0403 |
From: je...@comix.santa-cruz.ca.us (Jeff Liebermann)
Subject: Re: SCO Open Server VS Linux
Date: 1996/12/04
Message-ID: <583d5r$1cm@comix.comix.santa-cruz.ca.us>
X-Deja-AN: 202283235
references: <329F11FB.27E1@ix.netcom.com>
<57os2d$h0t@comix.comix.santa-cruz.ca.us> <E1rwI8.3ry@bigbird.telly.org>
organization: COmmittee to Maintain an Independent Xenix
newsgroups: comp.unix.sco.misc
Evan Leibovitch (e...@bigbird.telly.org) wrote:
: In article <57os2d$...@comix.comix.santa-cruz.ca.us>,
: Jeff Liebermann <je...@comix.santa-cruz.ca.us> wrote:
: This is all fine ans well, but does not answer the question one shred.
Sure it does. He wanted to know if anyone has tried Linux
and what we would suggest he install. Since he didn't
specify what he was going to do with the machine, I just
supplied my basic guideline for dealing with marginal decisions;
reduce the decision to dollars. Since there was no request
for relative technical merits, I supplied none.
: If there is any evidence whatsoever that one of the commercially-sold
: Linux distributions (Caldera or Red Hat) is less stable than SCO,
: let's hear it. So far all I've heard in this realm is FUD.
Warning: The following is pure and unadulterated FUD. When
customers use a better criteria for major business decisions,
I'll give up expousing it.
1. There is no measureable criteria for reliability in the computer
biz. Unlike benchmarks, claiming fault tolerance, crash proofness,
error recovery, MTBF, MTTF and uptime as a measureable indication
of reliability is ludicrous. Is a server with an average uptime of
12 months better than one that can only demonstrate 6 months? Nope.
Until a repeatable or demonstratable benchmark is established,
reliability remains empirical and very subjective.
2. ALL operating systems and small computers stink when it comes
to reliability. If they were so wonderful, nobody would need
Tandem non-stop kernels, mirroring, RAID, server clusters, live
backups and my services. Therefore, the question is not how reliable
is an operating system, but how many ways can it be destroyed
and what can be done about it. I recommend "Unix Haters Handbook"
by Garfinkel, Weise and Strassman for details. Your systems are
reliable because *YOU* are part of the equation. Left to themselves,
most Unix boxes will destroy themselves. Therefore the ultimate
measure of reliability is how well it works left to itself without
your tender loving care. On that basis, I would declare my Xenix
customers systems as the most reliable.
3. I stated that the criteria for judgement should be based upon
the classic FUD line "can you afford not to buy IBM"? This was
the line used by IBM in the PC/XT days when clones were garbage
and IBM compatibility was a real issue. Nothing has changed but
the names of the players. Todays bloated abominations have not
improved compatibility one iota. What has happened is that the
number of programmers, consultants and support personalities has
increased sufficiently to provide a recovery mechanism. IBM
compatibility is also no longer an issue. It's now an enormous
pile of acronyms manufactured by the computer industry in a
desperate attempt to prevent divergence due to exessive innovation.
This is one of the major benifits of name recognition which Linux
currently lacks in the business community.
4. To push FUD to its limit, I suggest that an operating system
decision should be similar to an insurance policy. You don't
buy insurance for any positive reason. You buy it because you
can't afford a loss or a crash. I can probably fabricate 100
reasons to use SCO Unix and maybe 98 of those apply to Linux.
If you were in business, and one crash costs much more than the
original OS, would you care to risk those two reasons? I've
given this choice to real, live, checkbook carrying customers.
The answer is always, I want to play it safe, therefore I'll
take SCO.
5. I don't have much experience with Linux. I have it, use it,
fight with it, and have a good feel for what it does. The biggest
advantage and also the biggest problem is the constant updates.
I have customers where I have to schedule downtime weeks in advance.
These customers worship uptime and consider a reboot a major
distraction. Many of them are going to SAS (Single Applications
Servers) simply to prevent downtime during updates. My customers
have no interest in seeing me constantly updating the kernel.
: There are reasons to choose SCO or Linux. Reliability is not one of
: these issues. Installed according to the vendor's instructions, either
: can run very well.
I have two types of customers. Those that have crashed and burned
and those that are about to crash and burn. The ones that have
crashed and burned never want to repeat the experience and will
do anything (exept spend money) to prevent a reputation. The
ones that will soon crash and burn are about to learn the merits
of quality hardware, conservative implimentations, cover-thy-ass,
and reliability. Reliability is the only issue after a meltdown.
: The main difference is that, since Linux includes source, it's easier
: to muck with a Linux installation if you want. But if reliability
: is an issue, who's going to go about mucking with the source?
Ahah. A concession that the operator may have some part in the
reliability equation. My high uptime servers tend to be locked
in a closet and away from anyone that owns a computer book or
a screwdriver. These boxes are the absolute most reliable servers
I've seen. Nobody touches them exept those that I know have a
clue. I have a large collection of operator error horror stories
for all operating systems. Mucking with the source is not even
an issue when nobody gets near the server.
: I have yet to install a Linux system that's 'primarily a workstation'.
: They're Web servers and POP servers and Samba servers and firewalls and
: routers.
If you have an operator on the console, it's a workstation. If
the bulk of the work is done remotely, it's a server. (This is
my definition). I run NT 4.0 server on my laptop. Is this a
server or workstation? I don't know as many Linux users as SCO
users, but my guess is that most of them use Linux as a workstation.
I will conceed that Linux also makes a good server.
: >I could go on, but methinks the distinction is evident. If
: >your application and customer can tolerate the workstation
: >methodology, then by all means run Linux. If your application
: >and server are expect to play continuously, methinks SCO
: >Unix would be a better choice.
: I don't understand the leap of logic here. Nowhere do you say that Linux
: is unsuitable as a server, nor that it is incapable of reliable operation.
Suitable for the purpose is largely guesswork here as the original
question offered little applications information. Lacking information,
I detailed the superficial distinctions between a server and a
workstation and suggested that it be considered in a decision.
I should have added that about 80% of all Unix and Unix like
operating systems are largely identical and that the server vs
workstation issue is only a small part of the remainder. I know
of people using Windoze 95 as a "server".
: How many of the world's web servers run on Linux? How many on SCO?
: Do any of those thousands upon thousands of Linux webmasters have any
: reliability problems using their systems as (primarily) servers? Is
: the Web world lining up to dump their Linux systems now that FastStart
: is here? Don't bet on it.
In computers, delivering a better mousetrap does not automagically
guarantee a mass exodus to the promised lan. I have customers
who will probably use their Xenix systems into the next century
for no better reason than it's been paid for. The Internet is
the current "growth" market. However, that's not where the $$$
appears to be. It's in the Intranet that the players are
concentrating. I really don't know how well Linux is doing for
corporate networking. My guess is not well because nobody wants
to deal with mixed systems unless there's a good reason. This
appears to be where NT is making its sales. I've noticed that
corporate users don't "dump" their OS's even if their totally
unreliable. If it's a commercial OS, they'll pound it to death
and hire an army of experts before giving up. Usually someone
can be found to make it work. However, on the Internet, I've
notices some sites are Linux this week, BSDI the next, SCO after
that, and so on. It's very strange.
: It is true that SCO has far, far more commercial applications available
: for it than Linux. But for those apps running on both platforms, such as
: ADABAS or WordPerfect or Wabi, is SCO demonstrably more robust?
No. I can't put a number on robustness or reliability. I can
only offer my personal impressions and experience as you have so
eloquently done in the comp.unix.unixware.misc newsgroup on a
similar issue. Given the choice, and the ability to replace
some disgusting parts and pieces, I can make SCO play reliably.
I'm not so sure about Linux. Since I am part of my customers
reliability equation, I'll pick the safest route.
: I usually agree with Jeff on most things, but I guess we'll have
: to agree to disagree on Linux's ability to cross the leap from
: marginal to functional.
No problem. Let's try an analogy. There are many auto manufacturers
producing many different types of vehicles. Each one will get you
from point A to point B. Lacking technical input, I can distingish
between models by size, performance and price. Reliability would
be difficult to determine without substantial history. Consumer
Report does this and gets it about 30% correct. By the time
sufficient history is accumulated to fabricate a decent review,
the auto is obsolete. The same with Unix. Linux is a sports
car with lots of hang on options. SCO Unix is 2.5 ton truck with
a high price and not much flash. The sports car can do really
neat and tricky things. The truck just gets you there and back.
The sports car needs lots of tuning but goes much faster and
runs more efficiently than the truck. The truck sucks resources
but will run for a million miles. Ad nasium. Each vehicle and
operating system has its place. When I say marginal or functional,
I mean marginal or function for the purpose intended (which is
currently unspecified).
: Compare OS5's non-Morningstar networking to
: Linux's out-of-the-box stuff and tell me, with a straight face,
: that this aspect of Linux is any less functional than SCO.
Touche. OSR5's networking is better than it's predicessor
but has a long way to go before it compares with Morningstar PPP.
I'll confess to limited experience with Linux PPP. Linux PPP
did work quite well after I read the HOWTO and corrected my
mistakes.
: I still say that the day Oracle and Informix some out with native
: ports for Linux, SCO has something real to worry about. The very
: high-end of SCO's market will be safe, but the low and middle
: ground will be very much up for grabs.
Have you seen the prices for Oracle and Informix database server
licences? I don't have the numbers handy but methinks they are
way out of line for the typical Linux customer. Why would Oracle
or Informix want to expand their installed base at the expense
licence fees? I see a challenge from one of the dbm workalikes,
but not from the major dbm vendors.
Back to the original question. The following is my not so humble
opinion of the best OS for the specific application for this week.
Apps OS
Oracle, Informix database server. SCO 3.2v5.0.x
Progress database server. SCO 3.2v4.2
Network applications development. Sun (Solaris)
Internet Web Server. BSDI
Data collection (SCADA) Flex OS (obsolete)
Character based vertical market apps. Xenix
MS Windoze applications server. NT 4.0
Real Time OS. QNX
911 emergency center, banking, transaction. Tandem Non-Stop
General purpose scientific workstation. Sun (Solaris)
General purpose productivity workstation. Windoze 95 or NT 4.0
Netware and Unix server mix. Unixware 2.1
Graphics and Web design. MacIntosh
Game machine. Windoze 95.
Ultimate geek and nerd workstation. Linux
Plug and Play computing Nintendo or Sega
[x] Email to author [ ] To mailing list [x] Posted to newsgroup
--
# Jeff Liebermann Liebermann Design 150 Felker St #D Santa Cruz CA 95060
# 408.336.2558 voice wb6ssy@ki6eh.#cenca.ca.usa wb6ssy.ampr.org 44.4.18.10
# 408.699.0483 digital_pager 73557,2074 cis [don't]
# je...@comix.santa-cruz.ca.us http://www.cruzio.com/~jeffl
From: da...@hendrix.postino.com (Danny Aldham)
Subject: Re: SCO Open Server VS Linux
Date: 1996/12/04
Message-ID: <584mh2$hbu@hendrix.postino.com>#1/1
X-Deja-AN: 202367463
references: <329F11FB.27E1@ix.netcom.com>
<57os2d$h0t@comix.comix.santa-cruz.ca.us> <E1rwI8.3ry@bigbird.telly.org>
organization: Postino DotCom
newsgroups: comp.unix.sco.misc
X-Newsreader: TIN [version 1.2 PL2]
Evan Leibovitch (e...@bigbird.telly.org) wrote:
: This is all fine ans well, but does not answer the question one shred.
: If there is any evidence whatsoever that one of the commercially-sold
: Linux distributions (Caldera or Red Hat) is less stable than SCO,
: let's hear it. So far all I've heard in this realm is FUD.
IF Hackers and denial of service count as less stable,
From a recent CERT advisory:
CERT(sm) Summary CS-96.6
November 26, 1996
Recent Activity
- ---------------
Since the September CERT Summary, we have noticed these continuing trends
in incidents reported to us.
1. cgi-bin/phf Exploits
2. Continuing Linux Exploits
We continue to see incidents in which Linux machines have been the victims
of root compromises. In many of these incidents, the compromised systems
were unpatched or misconfigured, and the intruders exploited well-known
vulnerabilities for which CERT advisories have been published.
If you are running Linux, we strongly urge you to keep current with all
security patches and workarounds. If your system has been root compromised,
we also recommend that you review
ftp://info.cert.org/pub/tech_tips/root_compromise
Further, you may want to monitor the Linux newsgroups and mailing lists for
security patches and workarounds. More information can be found at
http://bach.cis.temple.edu/linux/linux-security/
--
Danny Aldham www.postino.com
Need a UUCP E-Mail feed? - Ask me!
Dial-up access in British Columbia. TCP access around the World.
From: Mike Jagdis <m...@roan.co.uk>
Subject: Re: SCO Open Server VS Linux
Date: 1996/12/06
Message-ID: <E1zqI6.LIn@roan.co.uk>#1/1
X-Deja-AN: 202669768
x-mail2news-path: relay-11.mail.demon.net!relay-10.mail.demon.net!
roan.demon.co.uk!toaster.roan.co.uk
references: <329F11FB.27E1@ix.netcom.com>
<57os2d$h0t@comix.comix.santa-cruz.ca.us> <E1rwI8.3ry@bigbird.telly.org>
<584mh2$hbu@hendrix.postino.com>
x-mail2news-user: n...@toaster.roan.co.uk
organization: Roan Technology Ltd.
x-comment-to: da...@hendrix.postino.com (Danny Aldham)
newsgroups: comp.unix.sco.misc
Danny Aldham said
>IF Hackers and denial of service count as less stable,
>From a recent CERT advisory:
>[...]
It is worth noting that 99% of the Linux vulnerabilities occur because
of code which is *not* Linux specific but is widely used on many
other systems. There are large numbers of reported Linux attacks
because there are so many desktop Linux systems in use, many of them
in traditionally vulnerable places like universities and colleges.
There are large numbers of bugs reported on Linux rather than other
systems using the same code for two reasons: firstly because there
are many people actively reading and analysing 10+ year old code
to make Linux secure, secondly because CERT tends not to announce
bugs in commercial OS' until a fix is available. Just because you
don't know about the problem doesn't necessarily mean it isn't
there!
Mike
--
.----------------------------------------------------------------------.
| Mike Jagdis | Internet: mailto:m...@roan.co.uk |
| Roan Technology Ltd. | |
| 54A Peach Street, Wokingham | Telephone: +44 118 989 0403 |
From: Mike Jagdis <m...@roan.co.uk>
Subject: Re: SCO Open Server VS Linux
Date: 1996/12/06
Message-ID: <E1zq0y.LDG@roan.co.uk>
X-Deja-AN: 202669771
x-mail2news-path: relay-7.mail.demon.net!relay-5.mail.demon.net!
roan.demon.co.uk!toaster.roan.co.uk
references: <329F11FB.27E1@ix.netcom.com>
<57os2d$h0t@comix.comix.santa-cruz.ca.us> <E1rwI8.3ry@bigbird.telly.org>
<583d5r$1cm@comix.comix.santa-cruz.ca.us>
x-mail2news-user: n...@toaster.roan.co.uk
organization: Roan Technology Ltd.
x-comment-to: je...@comix.santa-cruz.ca.us (Jeff Liebermann)
newsgroups: comp.unix.sco.misc
Jeff Liebermann said
>1. There is no measureable criteria for reliability in the computer
>biz. Unlike benchmarks, claiming fault tolerance, crash proofness,
>error recovery, MTBF, MTTF and uptime as a measureable indication
>of reliability is ludicrous. Is a server with an average uptime of
>12 months better than one that can only demonstrate 6 months? Nope.
>Until a repeatable or demonstratable benchmark is established,
>reliability remains empirical and very subjective.
Reliability is measured using statistical analysis. All the above
are not unreasonable measures of reliability (although they generally
omit to mention the variance in the figures). Trying to treat
reliability as an absolute *is*, however, ludicrous.
>2. ALL operating systems and small computers stink when it comes
>to reliability. If they were so wonderful, nobody would need
>Tandem non-stop kernels, mirroring, RAID, server clusters, live
>backups and my services. Therefore, the question is not how reliable
>is an operating system, but how many ways can it be destroyed
>and what can be done about it. I recommend "Unix Haters Handbook"
>by Garfinkel, Weise and Strassman for details. Your systems are
>reliable because *YOU* are part of the equation. Left to themselves,
>most Unix boxes will destroy themselves. Therefore the ultimate
>measure of reliability is how well it works left to itself without
>your tender loving care. On that basis, I would declare my Xenix
>customers systems as the most reliable.
I have never seen a Unix box "destroy itself" when "left to itself".
At least, not once configured to do the job required of it. Of course,
the hardware sometimes intervenes in somewhat interesting ways on small
systems. High end systems provide high reliability hardware, intelligent
fail over, run time upgradable kernels etc. sure - but they are designed
for different requirements.
>4. To push FUD to its limit, I suggest that an operating system
>decision should be similar to an insurance policy. You don't
>buy insurance for any positive reason. You buy it because you
>can't afford a loss or a crash.
Sounds pretty positive to me. It's called "hedging your bets". You don't
buy insurance because you can't afford a loss or crash. You buy it in
order to create a positive return in such an event so as to minimise
the loss incurred in such an event. The amount of positive return you
choose is related both to how big a loss you would incur and how likely
you perceive such an event to be. It is just commercial risk analysis.
>I can probably fabricate 100
>reasons to use SCO Unix and maybe 98 of those apply to Linux.
>If you were in business, and one crash costs much more than the
>original OS, would you care to risk those two reasons?
Without having the faintest clue what these two reasons might be
how am I supposed to judge? It doesn't even qualify as a hypothetical
question. If you swap Linux and SCO in the above question what answer
do you get?
>I've given this choice to real, live, checkbook carrying customers.
>The answer is always, I want to play it safe, therefore I'll
>take SCO.
Precisely. but if you swapped Linxu and SCO they would say Linux
I bet. If the price difference is reasonably small and you tell
them that X is more reliable than Y they will naturally choose X.
If you actually ask this question of customers in this form then
you are deceiving your customers and misrepresenting both SCO
and Linux.
>
>5. I don't have much experience with Linux. I have it, use it,
>fight with it, and have a good feel for what it does.
There seems to be some conflict in the above?
>The biggest
>advantage and also the biggest problem is the constant updates.
>I have customers where I have to schedule downtime weeks in advance.
>These customers worship uptime and consider a reboot a major
>distraction. Many of them are going to SAS (Single Applications
>Servers) simply to prevent downtime during updates. My customers
>have no interest in seeing me constantly updating the kernel.
So don't do it. Do the same as you do with SCO. Run with what works
and only upgrade when you need a fix or feature. Do you install
*every* SLS for SCO the moment it is available? I don't. Nor do
I rush around upgrading all the Linux machines that are running
quite happily out in the field.
>I have two types of customers. Those that have crashed and burned
>and those that are about to crash and burn.
I do not allow my customers to crash and burn without my prior
permission :-). The systems I manage are set up to give me advance
warning of things going wrong. Only sudden and catastrophic
failure is a problem. Customers have the option of taking that
risk or buying *serious* hardware.
>If you have an operator on the console, it's a workstation. If
>the bulk of the work is done remotely, it's a server. (This is
>my definition). I run NT 4.0 server on my laptop. Is this a
>server or workstation? I don't know as many Linux users as SCO
>users, but my guess is that most of them use Linux as a workstation.
>I will conceed that Linux also makes a good server.
Your guess may be right if you are considering home systems but
very nearly *all* the systems I install - both SCO and Linux
are run as servers with no one using the console. the exception
is the Linux box on my desk which I run more or less as an X terminal
to give easy management of all the other machines. (The other
desktops are mostly dumb terminals and Win95 which gives 10x the
trouble of anything else used here :-) )
>If it's a commercial OS, they'll pound it to death
>and hire an army of experts before giving up. Usually someone
>can be found to make it work. However, on the Internet, I've
>notices some sites are Linux this week, BSDI the next, SCO after
>that, and so on. It's very strange.
Some people like to actually *examine* the different capabilities
instead of relying on heresay and guesswork. They will probably
experiment with free Unixware when it becomes available, they
would probably like to look at Solaris and probably wish they
could afford to play with Sparc, Alpha and MIPS boxes too.
It's only strange because you assume that *your* world view
applies to them. They would probably think you are strange for
*not* exploring the possibilities so as to provide best service
to your customers.
>No. I can't put a number on robustness or reliability. I can
>only offer my personal impressions and experience as you have so
>eloquently done in the comp.unix.unixware.misc newsgroup on a
>similar issue. Given the choice, and the ability to replace
>some disgusting parts and pieces, I can make SCO play reliably.
>I'm not so sure about Linux. Since I am part of my customers
>reliability equation, I'll pick the safest route.
"Reliability" again? If you could panic SCO from user space and
cause it to go down would you consider it to be more or less
reliable than Linux/BSD/Solaris/... If such a bug was known to
be being actively exploited how fast could you guarantee to
have your customers critical machines fixed to continue giving
the degree of service they required? (I used to have a trivial
little program that wiped out SCO 3.2.4 by opening a TLI TEP
as an ordinary user and doing a t_optmgmt with a buffer containing
bogus pointers.)
>Linux is a sports
>car with lots of hang on options. SCO Unix is 2.5 ton truck with
>a high price and not much flash.
The OS is more like the driver than the vehicle. Some driver are
quite happy behind the wheel of a removals lorry and drive a
car to work and back. Just occasionally you find an F1 driver
who also races karts and trucks (but knows when *not* to race as
well) and flies a helicopter. If you are friends with someone
like that and want to move house do you accept their offer to
drive a hire van for you or do you go with a professional company?
What do you base your decision on? The F1 driver because he can
also race single seaters? Or the professional because he will
charge lots for the job? Either way you would be best advised
to make sure you get the right vehicle for the job and get
some insurance for your possesions while in transit.
>Why would Oracle
>or Informix want to expand their installed base at the expense
>licence fees? I see a challenge from one of the dbm workalikes,
>but not from the major dbm vendors.
Because database systems are slowly moving to distributed object
paradigms. Oracle and/or Informix may decide that they need to
build a widespread presence on small systems to attract people
to their large systems and to provide a migration path that will
allow them to establish new technologies from the bottom up.
Similar reasons to those that lead to SCO providing FreeSCO to
the bottom end of the market.
At the moment many smaller systems are written and sold using
Sybase SQL Anywhere or MS Access simply because the entry point
for Oracle and Informix is *way* too high. It is easy to see
either of them introducing a low end product to encourage low
end development and awareness.
Mike
--
.----------------------------------------------------------------------.
| Mike Jagdis | Internet: mailto:m...@roan.co.uk |
| Roan Technology Ltd. | |
| 54A Peach Street, Wokingham | Telephone: +44 118 989 0403 |
From: je...@comix.santa-cruz.ca.us (Jeff Liebermann)
Subject: Re: SCO Open Server VS Linux
Date: 1996/12/07
Message-ID: <58b914$gbv@comix.comix.santa-cruz.ca.us>
X-Deja-AN: 202836227
references: <329F11FB.27E1@ix.netcom.com>
<57os2d$h0t@comix.comix.santa-cruz.ca.us> <E1rwI8.3ry@bigbird.telly.org>
<583d5r$1cm@comix.comix.santa-cruz.ca.us> <E1zq0y.LDG@roan.co.uk>
organization: COmmittee to Maintain an Independent Xenix
newsgroups: comp.unix.sco.misc
Mike Jagdis (m...@roan.co.uk) wrote:
I can tell this is going to be a hard sell...
: I have never seen a Unix box "destroy itself" when "left to itself".
: At least, not once configured to do the job required of it. Of course,
I have. Most of my business comes from following in the footsteps
of vertical market software suppliers and experts. I've seen:
1. Log file overflows
2. Disk space overflows
3. UPS's with 10 seconds of runtime left in the batteries.
4. Unscheduled reboots resulting in fsck trashing important files.
5. INN 1.4 and/or named hog memory until the system swaps.
6. Marginal hardware causes marginal reliability and marginal uptime.
7. Unverified tape backup that proves to be worthless for a restore.
8. Misconfigured system requires reboots to clear problems.
9. Floppy tape drives from hell.
10. Memory leaks.
11. User errors.
12. Minor update causes major headaches.
That should do. I can add more if anyone wants. These were systems
that were installed, configured and left to rot. Without preventive
maintenance, *ANY* Unix box will eventually slam into a problem of
some sort. IMHO, the service organization is a key component of
the overall reliability picture.
: the loss incurred in such an event. The amount of positive return you
: choose is related both to how big a loss you would incur and how likely
: you perceive such an event to be. It is just commercial risk analysis.
Sorry. I don't know the lingo. I buy insurance because the landlord
demands it and because a major disaster will prevent me from seperating
my customers from their money. I am willing to bet some of this money
with my insurance agent in trade for a promise that he'll come to my
rescue when I screw up.
As in buying insurance, purchasing an operating system works the
same way. Customers are willing to pay extra money for the dubious
promise that the system vendor will come to their rescue when
something goes amis. This is called support. The same customer
might be willing to pay even more for the dubious assertion that
one vendors product just might be a bit better and not require
as much support. After all, if these machines and operating systems
were so reliable, then there would be no market for backup devices,
mirroring, redunancy, hot standby, and support. Incidentally,
one of my customers actually bought downtime and crash recovery
insurance.
When the cost of a single crash and a day's downtime is equal to
the cost of a server, then paranoia, FUD, and conservative thinking
predominate the decision making process. The original question
was "which should I buy, SCO Unix or Linux?". I attempted to
explain the distinction between the use of an operating system
and how it fits in the decision making process. My customers are
medical billing offices, warehouses, wholesalers, hospitals,
and businesses that can usually lay claim to running "mission
critical" applications. Since I claim myself to be a major
component of the reliability picture, which these customers
demand, I tend to chose something that I know works and that I
know can be relied upon to do the job. As I said, I know I can
keep an SCO Unix server up and running reliably. I'm not so sure
about Linux.
: Precisely. but if you swapped Linxu and SCO they would say Linux
: I bet. If the price difference is reasonably small and you tell
: them that X is more reliable than Y they will naturally choose X.
: If you actually ask this question of customers in this form then
: you are deceiving your customers and misrepresenting both SCO
: and Linux.
I won't claim to have offered the choice fairly and impartially.
I will claim to have given everyone the opertunity to make the
switch. The issue was driven by the applications and database
vendors. They were abandoning Xenix and the various customers
had to chose something. In most cases, the decision was made by
the applications vendor without any consideration of alternatives.
You may find this hard to believe, but the cost of the OS was not
an issue. Who was going to keep it alive and how long was the
investment going to last were the major topics. Reliability crept
in at various points. If offering a preference to what I can
make work and keep alive constitutes deception on my part,
then I plead guilty.
: >5. I don't have much experience with Linux. I have it, use it,
: >fight with it, and have a good feel for what it does.
: There seems to be some conflict in the above?
Well, I've been submerged in Xenix since about 1986 (Xenix 2.0beta).
I've played with SCO Unix 3.2 since 1989. I've been tinkering with
OSR5 since mid 1995. At the same time, I have most of the Novell
and MicroSloth offering running. Sitting next to me are servers with:
SCO 3.2v4.2
SCO 3.2v5.0.2
SCO Xenix 2.3.4
Unixware 2.1.1
NT 4.0 Server SP1
Windoze 95 SP2
Linux 2.1.16 (I know I'm behind).
and behind me:
Novell 3.12
Novell 4.11
on the same lan. There is only so much time I can devote to
learning a new operating system. NT 4.0, OSR5, Linux and UW 2.1.1
were introduced in my office at the same time thus dividing my time.
However, comparisons and interoperabalony testing is now easy as
each of these runs on seperate servers. I will not claim expertise
with Linux. I will claim some experience.
: So don't do it. Do the same as you do with SCO. Run with what works
: and only upgrade when you need a fix or feature. Do you install
: *every* SLS for SCO the moment it is available? I don't. Nor do
: I rush around upgrading all the Linux machines that are running
: quite happily out in the field.
I'm not going to inumerate how many systems I take care of. In
order to reduce my workload and retain my sanity, I make a serious
attempt to keep all my servers running on the same version, with
identical updates and with identical versions of applications
software. I test the fixes and updates first in my office, and
then apply them to each and every customers machine. No exeptions.
It isn't just me that's applying updates. The applications vendors
are adding stuff as well as new versions of net software. At some
point, I have to catch up. Every 6 months is about right.
I'm also observing a trend towards single-applications servers.
These are boxes that run *ONE* application. A database server
would by a good example. Same with a Lotus Notes server. A
web server might be considered single-application. Definately,
a news server. What suprised me was that this trend was appearing
in some of the larger corporate server closets. The MIS mob said
it was to reduce vulnerability to downtime caused by various
legitimate circumstances. If you have a swiss-army-knife server
that runs a multitude of appications, the chances that a bad update
to one application will down the server and prevent access to
others that are working perfectly. These MIS guys have been there
and are apparently on the defensive. No clue how they handle
updates (I didn't ask).
: I do not allow my customers to crash and burn without my prior
: permission :-). The systems I manage are set up to give me advance
: warning of things going wrong. Only sudden and catastrophic
: failure is a problem. Customers have the option of taking that
: risk or buying *serious* hardware.
I don't have that luxury. I tend to inherit systems rather than
sell them. Upgrades tend to be evolutionary. One customer still
has 4 generations of previous servers as a result of paranoia over
an upgrade. I monitor diskspace, hung processes, run-away process,
cron jobs, backups, etc. My pager goes off at inconvenient times.
However, there's always a new problem to keep me entertained.
There's nothing better than a regular physical inspection and tour
of the filesystems.
: Some people like to actually *examine* the different capabilities
: instead of relying on heresay and guesswork. They will probably
: experiment with free Unixware when it becomes available, they
: would probably like to look at Solaris and probably wish they
: could afford to play with Sparc, Alpha and MIPS boxes too.
You mean they don't just find something and make it work?
Perhaps they're not happy with what they have and are desperately
looking for an alternative? I would wonder why a commercial
ISP could rotate operating systems on his production server.
Switching is lots of work, switching back is even more, and
the risks are enormous. If I start to see musical operating
systems, I start to worry. The smart ones run alternative
systems on various offline servers and then deploy them after
some testing.
: It's only strange because you assume that *your* world view
: applies to them.
Well, you have a point. I have definite opinions and have little
trouble propogating them. See next item.
: They would probably think you are strange for
: *not* exploring the possibilities so as to provide best service
: to your customers.
That's a tough one. The cost of experimentation is high.
Dealing with an unknown is far more risky than dealing with
a known problem. Some ISP's engage in some of the strangest
rituals in the name of uptime and continued service. A new
OS could either be the ultimate solution or the biggest
nightmare. The problem is that customers are paying for the
service and might not find a botched operating system transplant
particularly entertaining. I'll leave it to the ISP to weigh
the risks. I can't tell from here.
: "Reliability" again? If you could panic SCO from user space and
: cause it to go down would you consider it to be more or less
: reliable than Linux/BSD/Solaris/...
That's easy. If *I* could panic it, I would be worried but not
much more. If *my customer* could panic it, it would be a
candidate for replacement. If the OS could panic itself, it
ends up in the dumpster. Strangely enough, my customer base
does very little on the command line and runs almost exclusively
applications software from menus. I have a few denial of service
attacks that will busy out any system. This doesn't worry me.
Security issues are a real problem but not from the standpoint
of hacking. It's access by former employees, accidental erasures,
and idiot errors that worry me.
: If such a bug was known to
: be being actively exploited how fast could you guarantee to
: have your customers critical machines fixed to continue giving
: the degree of service they required?
I would expend more energy finding the culprit and arranging to
have the bozo fired. If it came from an applications program,
the software vendor will have to do some explaining. I've had
situations like this before and found it easier to fix the cause
of the problem instead of doing damage control on the OS. That
can come later.
: >Why would Oracle
: >or Informix want to expand their installed base at the expense
: >licence fees? I see a challenge from one of the dbm workalikes,
: >but not from the major dbm vendors.
: At the moment many smaller systems are written and sold using
: Sybase SQL Anywhere or MS Access simply because the entry point
: for Oracle and Informix is *way* too high. It is easy to see
: either of them introducing a low end product to encourage low
: end development and awareness.
There was an article in PC Week 2 weeks ago on this issue.
If Oracle and Informix wanted an entry level version, they would
grade their offerings by user count with appropriate user bumps.
In theory, there should be no price distinction between an
OSR5 and a Linux version. However, let's assume that they
actually do scale the entry price by operating system. When
Linux finally has a track record and sufficient history to
determine its reliability, what's there to prevent them from
reclassifying Linux as a mainstream OS worthy of predatory
pricing?
All your comments are well thought out and show experience.
What should be obvious is that there are many different
applications, customers, configurations and uses for an
operating system. Blanket generalizations just don't work.
Tell me what you're going to do with Unix or Linux and
in what kind of environment, and then we can talk about
suitability, reliability, support, FUD and superiority.
[x] Email to author [ ] To mailing list [x] Posted to newsgroup
--
# Jeff Liebermann Liebermann Design 150 Felker St #D Santa Cruz CA 95060
# 408.336.2558 voice wb6ssy@ki6eh.#cenca.ca.usa wb6ssy.ampr.org 44.4.18.10
# 408.699.0483 digital_pager 73557,2074 cis [don't]
# je...@comix.santa-cruz.ca.us http://www.cruzio.com/~jeffl
From: j...@ns.cybersurf.co.uk
Subject: Re: SCO Open Server VS Linux
Date: 1997/02/21
Message-ID: <ufn2sytqp2.fsf@ns.cybersurf.co.uk>#1/1
X-Deja-AN: 220380892
References: <329F11FB.27E1@ix.netcom.com> <57n70e$khj@nntp.interaccess.com>
Organization: Cable Internet
Newsgroups: comp.unix.sco.misc
I like Linux (a lot) but, as systems administrator of a large web authoring /
hosting business, I run SCO on any UNIX PC boxes we have. The simple reason
is the stigma attached (rightly or wrongly is irrelevant) to "free" software.
Admittedly, things are beginning to turn. We are seeing a lot of smaller
companies getting connected, and almost all of them go for Linux, but
personally *much though I admire linux for it's performance and usability*
I prefer to use a system with an established track record.
John
From: Paul Anderson <ander...@agapesystems.com>
Subject: Re: SCO Open Server VS Linux
Date: 1997/02/22
Message-ID: <330F9367.63DE@agapesystems.com>#1/1
X-Deja-AN: 221502107
References: <329F11FB.27E1@ix.netcom.com> <57n70e$khj@nntp.interaccess.com>
To: j...@ns.cybersurf.co.uk
Organization: Agape Information Systems
Newsgroups: comp.unix.sco.misc
j...@ns.cybersurf.co.uk wrote:
>
> I like Linux (a lot) but, as systems administrator of a large web authoring /
> hosting business, I run SCO on any UNIX PC boxes we have. The simple reason
> is the stigma attached (rightly or wrongly is irrelevant) to "free" software.
>
> Admittedly, things are beginning to turn. We are seeing a lot of smaller
> companies getting connected, and almost all of them go for Linux, but
> personally *much though I admire linux for it's performance and usability*
> I prefer to use a system with an established track record.
>
> John
I have both SCO and Linux. In 1990 Linux was introduced for a project
in which we had to write an embedded Unix for a receiver. I passed over
it with the "free" software stigma. Linux has improved at a very rapid
rate. All the interfaces are grown up and mature and the OS is matuer
enough to have won a desktop software of the year award.
SCO is getting no such accolades for their old tired 5.0 product. Their
IXI desltop is good their OS is just plain old. It took SCO forever to
move off 5.2, now 5.3. It does have some in roads but is very dated in
many regards. The first that comes to mind is the serial drivers and
uucp. And that's just for starters. SCO owns rights to Unix, and Bill
owns 35% of SCO, so I'm not expecting much to change. Bill would love
to see Unix die.
Paul Anderson
From: bal...@world.std.com (Jim Balson)
Subject: Re: SCO Open Server VS Linux
Date: 1997/02/26
Message-ID: <E68Cn3.533@world.std.com>#1/1
X-Deja-AN: 221652322
References: <329F11FB.27E1@ix.netcom.com> <57n70e$khj@nntp.interaccess.com>
<330F9367.63DE@agapesystems.com>
Organization: The World Public Access UNIX, Brookline, MA
Newsgroups: comp.unix.sco.misc
Paul Anderson (ander...@agapesystems.com) wrote:
: j...@ns.cybersurf.co.uk wrote:
: >
: > I like Linux (a lot) but, as systems administrator of a large web authoring /
: > hosting business, I run SCO on any UNIX PC boxes we have. The simple reason
: > is the stigma attached (rightly or wrongly is irrelevant) to "free" software.
: >
: > Admittedly, things are beginning to turn. We are seeing a lot of smaller
: > companies getting connected, and almost all of them go for Linux, but
: > personally *much though I admire linux for it's performance and usability*
: > I prefer to use a system with an established track record.
: >
: > John
: I have both SCO and Linux. In 1990 Linux was introduced for a project
: in which we had to write an embedded Unix for a receiver. I passed over
: it with the "free" software stigma. Linux has improved at a very rapid
: rate. All the interfaces are grown up and mature and the OS is matuer
: enough to have won a desktop software of the year award.
: SCO is getting no such accolades for their old tired 5.0 product. Their
: IXI desltop is good their OS is just plain old. It took SCO forever to
: move off 5.2, now 5.3. It does have some in roads but is very dated in
: many regards. The first that comes to mind is the serial drivers and
: uucp. And that's just for starters. SCO owns rights to Unix, and Bill
: owns 35% of SCO, so I'm not expecting much to change. Bill would love
: to see Unix die.
: Paul Anderson
No arguments about the comments on SCO Openserver. It is old and tired.
But wait till SCO releases "Free Unixware" within the next few
weeks or so. I'll take "Free Unixware" over Linux anyday, especially since
the next release of Unixware, due out this summer/fall, will have CDE
attached to it. CDE, being a $250 option for Linux, will be free with
Unixware.
Jim
--
Jim Balson
bal...@world.std.com
From: e...@bigbird.telly.org (Evan Leibovitch)
Subject: Re: SCO Open Server VS Linux
Date: 1997/02/28
Message-ID: <E6AsqF.Jzt@bigbird.telly.org>#1/1
X-Deja-AN: 222021776
References: <329F11FB.27E1@ix.netcom.com> <57n70e$khj@nntp.interaccess.com>
<330F9367.63DE@agapesystems.com> <E68Cn3.533@world.std.com>
Organization: Sound Software
Newsgroups: comp.unix.sco.misc
In article <E68Cn3....@world.std.com>, Jim Balson <bal...@world.std.com> wrote:
>Paul Anderson (ander...@agapesystems.com) wrote:
>: > We are seeing a lot of smaller
>: > companies getting connected, and almost all of them go for Linux [...]
>: I have both SCO and Linux. In 1990 Linux was introduced for a project
>: in which we had to write an embedded Unix for a receiver.
>But wait till SCO releases "Free Unixware" within the next few
>weeks or so. I'll take "Free Unixware" over Linux anyday, especially since
>the next release of Unixware, due out this summer/fall, will have CDE
>attached to it. CDE, being a $250 option for Linux, will be free with
>Unixware.
Jim, all of the posting on this issue for the past weeks have centered
on the changing level of acceptability of Linux in *business* environments.
Free UnixWare, no matter what bells and whistles are hung from it, will
not be legal for use in any of these situations. Those places will still
have to shell out $1,300+ to get their UnixWare and CDE.
--
Evan Leibovitch, Sound Software Ltd, located in beautiful Brampton, Ontario
Supporting PC-based Unix since 1985 / Caldera & SCO authorized / www.telly.org
Q: How many Microsoft employees does it take to change a lightbulb?
A: None. They merely declare darkness to be the new industry standard.