Re: Dell stops offering Red Hat on desktops

[Steven Pritchard <steve@silug.org>]

Tony Zafiropoulos said:
> I suppose this is momentum as well.  We need to show the gaming community
> how big the installed base is.

That's kind of a chicken-and-egg problem.  The installed base would be
a hell of a lot bigger if the drivers were better.  At least we have
lots of games now...

[about schools]
> This could be done with peer pressure and case studies, show them how much
> money they are saving. etc. etc.
[...]
> You need to put a case study together.

I guess this would be another chicken-and-egg thing, for me at least.
While I've seen a school use Linux extensively, I was not actually
involved.  I guess I really should find a local school that I can
apply gentle pressure to so I actually have those kinds of numbers...

I wonder if anybody has done this already though...

[financial software]
> We need examples, the ones that I am familiar with (small companies) like
> ILX and Reuters are all Windows platforms.

Got me...  That's just not my area of expertise.  (And I was thinking
more along the lines of Peachtree, Quickbooks, etc.)

> Ok, I received a MS packet "Competing with Linux" packet from a friend of
> mine. He just received this (June CD stamps).
[...]
> This is what they emphasized:
> 
> Against the "Linux is inexpensive" argument...  (The quoted Linux value
> proposition and the MS response)
> 
> Linux is definitely cheap, and even free if you are willing to download it
> yourself, but you get what you pay for. The cost of software is a small
> component of deploying and managing an enterprise IT infrastructure.

Wow, that's amazing.  You'd think that a company the size of Microsoft
would learn that half-truths and innuendo will only go so far.  (That
last sentence is true, BTW.)

> Microsoft software is already essentially "free" in large
> projects-typically less than 3% of the hardware/software cost and less
> than 1% of the total project cost.

Hmm...  If a project costs $1M, that makes the cost of Microsoft
software somewhere between $10k and $30k, even if you believe their
numbers (and I really don't).  Now, in exactly who's world is $30k an
insignificant amount of money?

Of course, I could always go into the old argument that it doesn't
take as much hardware to run a non-M$ solution, so the total cost of
the project is likely to be much lower to end up with the same
capabilities...

> Because supportability and integration are so important, Microsoft
> software and commercially supported software in general help customers
> where it counts: minimizing total cost of ownership and maximizing
> business value, making it less expensive for customers over the long haul.

Wow, and it only took how many years for Microsoft to figure this out?

[...]
> Microsoft products provide a rich set of third party applications and are
> backed by a large number of certified partners and support professionals
> like yourself.

"Do we need to remind you where your money comes from?"

> While most substantial Microsoft product offerings can never be without
> cost, the old saying about "getting what you pay for" certainly applies in
> this case.

OK, so now we've moved up to cliches and innuendo...

> "Linux is Open Source Code"
> 
> Customers want to have some influence over the direction of their
> computing platform. The lack of a single owner and well-known decision
> making process for open source software makes it difficult for customers
> to influence or guide the direction of features. There is no certainty
> that in the long term Open Source Software will evolve to meet the
> changing needs of the customer and the marketplace.

Implying, of course, that there is some kind of guarantee that
proprietary software "will evolve to meet the changing needs of the
customer".  Yeah, right.  How long did it take for Microsoft to
include networking support in Windows?  How long after that did it
take for them to include TCP/IP?  How long after that did it take for
them to include any basic TCP/IP services?  (Hint: They only did that
when their monopoly was threatened.)

Of course, the difference with Open Source is that a user is free to
make the changes on their own, or pay someone qualified to make the
changes.  With proprietary software, you get what the author gives
you.  Period.

> The open source development model also depends on its users for
> testing. Windows NT is tested daily on hundreds of OEM configurations,
> connected to thousands of peripherals, running thousands of different
> applications.

Uh-huh.  No Linux distribution vendors do any testing.  Nor do
hardware vendors.  I guess I'm not really seeing posts to linux-kernel
from redhat.com, ibm.com, sgi.com, valinux.com, ...

Oh, and guess what?  People testing on Linux can actually *fix* things
that are broken!

> Such a testing lab requires deep investments and strict discipline that
> cannot be easily reproduced through a loosely connected group of
> individuals on the Internet. 

Uh-huh, again.  I guess I haven't really seen all the detailed reports
on various kernel issues on linux-kernel, and I guess I imagined all
that automated testing software out there...

Besides, as Linus said, "Given enough eyeballs, all bugs are shallow."

> "Linux is more reliable"
> 
> Linux is being used for simple tasks such as file/print and static web
> page serving. Microsoft customers are using Windows NT Server for
> demanding, high performance, mission critical applications such as
> messaging, data warehousing, decision support and e-commerce. Less heavily
> loaded systems with less complex software suites have high reliability.

Yeah, right.  In a former life, I played NT admin.  Not one of the NT
boxes that I admin'd was the slightest bit stable, even compared to
the SCO boxes at the same company.  (I personally think SCO is the
worst excuse for a Unix that I've ever seen, if that tells you
anything.)

As an example, one server played file server and primary domain
controller for a couple hundred systems.  It would often die in the
middle of a work day for no apparent reason.  We were told that we
shouldn't have a domain controller doing anything else...

I could go on, but the point is that not *one* of the NT systems was
reliable.  The most reliable one we had was a NT 3.51 box that was
nothing but a print spooler for network printers.  It could actually
stay up for a few months at a time.

Contrast this to the Linux boxes, several of which were never rebooted
except when we lost power for extended periods of time.  One actually
ran for weeks at a time on a box that SCO refused to even boot on and
NT died on after a few minutes.

Of course, it's not just Linux that has this kind of high
reliability...  It's pretty standard for commercial Unix systems too. 
I've seen well over a year uptimes on HP-UX systems.  One went 700
days before an extended power outage forced a shutdown.

> Four of Microsoft's key OEMs (IBM, Compaq, HP and Data General) now offer
> 99.9% uptime guarantees for Windows NT Server, attesting to the high
> reliability of Windows NT.

And the commercial UNIX vendors are into the "5 Nines" (99.999%)
thing now...  That's a little over 5 minutes of downtime per *year*.
99.9% uptime translates to more like 9 hours of downtime per year.

Oh, and I think they do some tricks with failover and such just to get
to that point...

> Research on SP4 shows that a majority of
> Microsoft's customers see Windows NT Server as reliable or more reliable
> than either Netware or UNIX.

Research shows that a majority of people believe in aliens, think
Elvis is still alive, etc.  I bet I can even come up with real scientific
studies to back up my worthless statements too.

> Our actual operational uptime for a cadre of over 1100 servers in ITG with
> a wide mixture of workloads for a recent six-week period was 99.91%, and
> almost none of these systems were clustered. Clearly Windows NT is highly
> reliable when operated competently on good quality hardware and properly
> maintained with hot fixes and service packs.

There are Linux clusters ten times that size.  I wonder what sort of
average uptime they get...  Probably better than an hour of downtime
in a 6-week period.  (Do the math.  It's actually pretty pitiful.)

> "Linux has an extensive developer network"
> 
> Linux often uses the catch phrase "built by users for users" but a more
> realistic restatement is "built by developers for developers." The Linux
> development community is comprised of technical hobbyists and UNIX
> enthusiasts whose idea of usability is a good text editor with which to
> modify configuration files.

We're not at all biased here, are we?

Considering how big of a factor shareware was in the early popularity
of Windows, I find it more than a little amusing that Microsoft
refuses to see that users and developers are not necessarily distinct
groups.

> If the users are the developers, then the product will be shaped to suit
> the mind of a developer, with ease of use taking a back seat to
> flexibility. This developer focus also manifests itself in the way a
> problem is often resolved: posting a patch file that the user must apply
> to the original source file, then recompile and link it, and finally
> re-install it.

If the developers are the users, then the product will be shaped to
suit the mind of a user, with monopoly maintenance taking a back seat
to functionality.

And you know, I could have swore that the Linux distributions all had
spiffy easy ways of downloading and installing the latest updates.  I
didn't know you had to recompile everything each time.  (Is the
sarcasm getting to heavy for anyone else here?)

OK, I really have to get to the meeting now...  :-)

Steve
-- 
steve@silug.org           | Southern Illinois Linux Users Group
(618)398-7360             | See web site for meeting details.
Steven Pritchard          | http://www.silug.org/
-
To unsubscribe, send email to majordomo@silug.org with
"unsubscribe silug-discuss" in the body.