OEM vs Retail XP Pro

[Matt]

I found one story that claimed Linux had less than .25% marketshare on the
desktop worldwide, but it's based on browser hits:

http://www.macworld.com/news/2001/12/19/linux/

That's almost three and a half years old.

 

[Matt]

What recent report? Who sponsored the study? How did they determine this
crap?

Ruel, try this instead:
JANUARY 31, 2005 Business Week COVER STORY
Linux Inc.
http://www.businessweek.com/magazin...mpaign_id=nws_insdr_jan21&link_position=link1

 

[Mxsmanic]

Your logs.

Yes, from my Web site. I get about 300,000 unique visitors per month,
roughly a couple million individual page hits.

 

[Mxsmanic]

What recent report? Who sponsored the study? How did they determine this
crap?

These come off my logs. My site gets a visitors from a pretty broad
cross-section of home and office desktops.

Linux definitely has more marketshare than .3% on the desktop.

How do you know?

I run into
Linux users now and then, most of them are so geeky they're easy to spot,
but I've never run into a Mac user that wasn't shopping in the Mac section
at either Micro Center or CompUSA.

How is this relevant?

 

[Mxsmanic]

Well, either those numbers are a) fictitious b) from a study sponsored by MS
c) from a study based on browser hits d) obtained from someone who got them
from someone else and so on, or e) wishful thinking.

They are based on browser identifications.

I found one story that claimed Linux had less than .25% marketshare on the
desktop worldwide, but it's based on browser hits:

And what's wrong with that?

Here's another study that claims Linux has 3.2% marketshare in 2003, and
that it's higher than the Mac's:

http://www.itfacts.biz/index.php?id=P723

And where did they get their numbers?

There are conflicting marketshare studies all over the net. Like I've said
before, I've read that it could be as high as 13%.

Yes, and I may already be a winner of Publisher's Clearinghouse
Sweepstakes.

That figure was considering solo Linux boxes, dual boot machines, and
LiveCD (like Knoppix) ran Linux deployments.

And where did those numbers come from?

That means, most of those same users also run Windows.

How much time do they spend running Windows, and how much time do they
spend running Linux?

That makes it difficult to get a true figure on just how many
people are running Linux even on an occasional basis.

I agree.

If Linux had only a .3% worldwide marketshare on the desktop, the chances
of me running into a Linux user would be slim, let alone running into one
somewhere at a gym working out lifting weights.

Not really. That's still 1 in 300.

That figure has to be complete FUD.

Because you say so? What about the figures for the Macintosh and
Windows? Are they FUD as well? If not, why not?

 

[Mxsmanic]

That would make it 50% more popular than the Mac, which is pure fantasy,
even in 2007.

No, the Mac is the strongest rival that Windows has ever faced.

In servers, researcher IDC predicts Linux' market share based
on unit sales will rise from 24% today to 33% in 2007, compared
with 59% for Windows -- essentially keeping Microsoft at its
current market share for the next three years and squeezing
its profit margins.

It's not about sales, it's about users.

 

[Mxsmanic]

Please give evidence of a version of Windows that ran on a CPU that
didn't support the Intel instuction set.

Windows NT.

 

[John Doe]

I've read about some horror stories about when XP is installed
there might be a big hassle "reactivating??" it when one changes
major components such as motherboards, etc.--too big a monopoly
in my opinion.

WPA_Kill.exe

A little freedom Americans can re-import.

 

[John Doe]

Ruel Smith writes:

Interesting. Each new release of Windows wastes more and more
space on the desktop. But that seems to be true for a lot of
GUIs.

Windows XP looks exactly like Windows 99 here, with a few
exceptions that don't require screen space.

Currently, the only non-program window stuff on my main monitor is a
single row taskbar which includes two quick launch icons, program
buttons, and five system tray icons.

Multiple monitors increases screen space dramatically. My secondary
monitor has three program Windows and that's all.

Speech control eliminates the need for program icons.

 

[David Maynard]

David Maynard writes:




I know that, and you know that, but a lot of people don't know that.

Hehe. Well, those folks aren't going to be familiar with the command line
either and it's easier to teach someone "don't use this server as a
workstation" than it is to teach Linux and the command line.

Well, you do on Windows.

Well, you snipped the context but, no, you don't. Network access is network
access and does not 'use the GUI'.

That's one of the big handicaps of Windows for
server environments. There is just no way to administer Windows servers
with just a CLI.

Administration is a different matter than 'network data base access'.

The GUI is there to make it 'simple' to administer but you're not sitting
there 'administering' the thing 24/7 (or else you have bigger problems than
worrying about the GUI) so the GUI load is irrelevant 99% of the time.

Now, I'm not saying that is necessarily the 'correct' solution for a server
but that's Microsoft's basic philosophy: to, stealing a phrase, make it "so
simple even a caveman could do it."

 

[David Maynard]

Please give evidence of a version of Windows that ran on a CPU that
didn't support the Intel instuction set.

The initial release of Windows NT supported the x86 and MIPS architecture.
Support for the Digital Equipment Corporation (DEC) Alpha AXP was added
shortly thereafter. Support for a fourth processor architecture, the
Motorola PowerPC, was added in Windows NT 3.51.

By the time Windows 2000 rolled out there was little incentive to support
those any more, other than x86, and they were dropped.

 

[Matt]

Matt writes:




Yes, from my Web site. I get about 300,000 unique visitors per month,
roughly a couple million individual page hits.

By all means, mention your results to IDC. I expect they will be
interested to know that some anonymous webmaster contradicts their
estimates by a factor of 10.

 

[Ruel Smith]

Whoa, Ruel. I hope you aren't blaming me for
Maniac^H^H^H^H^H^HMxsmanic's claims.

No, I just didn't his message, but only your reply. Therefore, it was all I
could do to quote it through your message. I have no idea how his message
was lost.

 

[Ruel Smith]

They are based on browser identifications.

Which, again, are totally false, and I showed you why. You can't base it on
browser identification because Konqueror can appear as another browser on
another operating system altogether. You can't base it on sales because
most Linux installations are from totally free places like Debian, Gentoo,
Slackware, and even free versions of SuSE, Fedora (Red Hat), and Mandrake.
Also, that would be accounting for many computers that are sold with a
Windows license that get wiped in place of Linux. There simply is no way to
accurately measure Linux marketshare.

And what's wrong with that?

I've showed you before, and I'll show you again:

http://home.fuse.net/ruelsmith/SnapshotBrowserIdent.png

That's my system with Konqueror open and the menus dropped to show you where
you can set Konqueror to identify itself as Internet Explorer running on
Windows. It even lets you identify yourself as a number of versions of IE
running on different versions of Windows and even the Mac. It'll also allow
you to be identified as running Mozilla, Opera, Safari, or Netscape
Navigator on Windows or Mac as well.

Now, unless you're in total denial, you can see why browser hits isn't an
accurate measurement at all.

And where did they get their numbers?

Don't know, and don't care. Again, an accurate measurement isn't possible at
all.

Yes, and I may already be a winner of Publisher's Clearinghouse
Sweepstakes.

Congratulations... Bill will be pleased to get more of your money, since you
have more to give him...

And where did those numbers come from?

Yes, but how often? If 10% of the population runs Linux 50% of the time and
Windows 50% of the time, are they counted as Linux user or Windows user? I
have a Windows machine, but spend about 10% of my time on it. Should I be
counted as a Windows statistic _and_ a Linux statistic, or should each
person count once?

How much time do they spend running Windows, and how much time do they
spend running Linux?


I agree.


Not really. That's still 1 in 300.

No, 3 tenths of 1% is equal to 1 in 3000. Regardless, I run into Linux users
much more frequently than Mac users. You'd think that if the Mac had a
larger share of the market, I'd run into more of them instead.

Because you say so? What about the figures for the Macintosh and
Windows? Are they FUD as well? If not, why not?

Show me a figure that is accurate and prove to me how it is accurate and
I'll believe it. Until then, you're full of it.

 

[Ruel Smith]

That's almost three and a half years old.

Exactly. Mr. Mxsmanic claims that it only has .3% marketshare. A poor
estimate of .25% marketshare was estimated almost 3 and a half years ago.
It shows just how full of it that estimate is.

 

[Mxsmanic]

Well, you snipped the context but, no, you don't. Network access is network
access and does not 'use the GUI'.

You can't administer a Windows server without a GUI.

 

[JAD]

Lets see, I have a Mac running MAC OS, a MAC server running A/UX, 2
PC's 1 XP 1 ME., and a Linux box. Now the Linux box doesn't boot it just
there to beat up every time my other boxes don't run.(just my way of venting
frustration that its never helped in anyway to save the planet) Anyway, I
can't even think in the picture of stats where that would figure in.
Statistics are, as you have demonstrated, fuzzy misinformation givers on a
good day. You would have to ask the server for more room to post the
exceptions and rules of information acquisition..

 

[Matt]

0.3 % = 0.3 * 0.01 = 0.003 ~= 1/333

 

[Mxsmanic]

Which, again, are totally false, and I showed you why. You can't base it on
browser identification because Konqueror can appear as another browser on
another operating system altogether.

If there are millions of Linux users out there who deliberately choose
to pretend to be something else, that's their problem, not mine. It
requires quite a leap of faith to assume that so many users who appear
to be using Windows or the Mac are in fact Linux users in the total
absence of evidence. Prove to me that they are actually running Linux,
and I'll believe it, but I won't accept it as an article of faith.

You can't base it on sales because
most Linux installations are from totally free places like Debian, Gentoo,
Slackware, and even free versions of SuSE, Fedora (Red Hat), and Mandrake.

Endless millions of copies of Windows are pirated, too, and so they
don't count in sales, either, especially in Asia.

Also, that would be accounting for many computers that are sold with a
Windows license that get wiped in place of Linux. There simply is no way to
accurately measure Linux marketshare.

Which means there is no way to support your assertion that Linux has a
larger market share than browser statistics would indicate.

That's my system with Konqueror open and the menus dropped to show you where
you can set Konqueror to identify itself as Internet Explorer running on
Windows. It even lets you identify yourself as a number of versions of IE
running on different versions of Windows and even the Mac. It'll also allow
you to be identified as running Mozilla, Opera, Safari, or Netscape
Navigator on Windows or Mac as well.

Maybe Windows and Mac users are identifying themselves as Linux, too,
which would artificially _inflate_ the Linux figures.

Now, unless you're in total denial, you can see why browser hits isn't an
accurate measurement at all.

There's nothing more reliable at the moment.

Don't know, and don't care. Again, an accurate measurement isn't possible at
all.

So there's no reason to not believe browser numbers, because nothing
else is available.

Yes, but how often? If 10% of the population runs Linux 50% of the time and
Windows 50% of the time, are they counted as Linux user or Windows user? I
have a Windows machine, but spend about 10% of my time on it. Should I be
counted as a Windows statistic _and_ a Linux statistic, or should each
person count once?

Unless we have accurate numbers on how much time they spend with each
OS, there is no way to say. Nor is there any reason to speculate that
Linux somehow is more widely used than the available numbers imply.

No, 3 tenths of 1% is equal to 1 in 3000.

No, three tenths of 1% is equal to one in 300. 1/0.003 = 333.

I think I'm beginning to see the problem.

Regardless, I run into Linux users much more frequently than Mac users.

The people you run into are a much less reliable measure of Linux
penetration than the browser numbers in my logs, which are not subject
to bias.

You'd think that if the Mac had a larger share of the market, I'd
run into more of them instead.

That depends on the company you keep.

Show me a figure that is accurate and prove to me how it is accurate and
I'll believe it. Until then, you're full of it.

I'll take my browser figures over your circle of friends.

 

[John Doe]

Ruel Smith writes:

If there are millions of Linux users out there who deliberately
choose to pretend to be something else, that's their problem,
not mine. It requires quite a leap of faith to assume that so
many users who appear to be using Windows or the Mac are in fact
Linux users in the total absence of evidence. Prove to me that
they are actually running Linux, and I'll believe it, but I
won't accept it as an article of faith.

You can figure Linux users are more likely (than Windows users)
skilled and maybe motivated enough to do that.

That's hardly conclusive, it's just the first related argument
that came to my mind.

Have fun anyway.