Linked by David Adams on Mon 16th May 2011 02:30 UTC, submitted by Peter Alguacil
Linux Linux currently powers a majority of the world’s web servers and supercomputers. As a desktop OS, however, Linux has yet to gain mainstream acceptance. That said, there are some countries where people have embraced Linux on the desktop to a greater degree than most.
Order by: Score:
Numeric discrepancy
by lemur2 on Mon 16th May 2011 03:09 UTC
lemur2
Member since:
2007-02-17

FTA: "there are at least 15 million desktop Linux users out there".

I suppose that is tue enough, but there are 52 million Linux desktops deployed to students in Brazil alone.

http://aseigo.blogspot.com/2008/04/deploying-kde-to-52-million-youn...

http://desktoplinux.com/news/NS9272932512.html

Brazil somehow doesn't even appear in the list from the article.

Reply Score: 11

RE: Numeric discrepancy
by reldruh on Mon 16th May 2011 03:35 UTC in reply to "Numeric discrepancy"
reldruh Member since:
2007-02-05

You might want to read those articles again.

By the end of this year 29,000 labs serving some 32,000,000 students will be fully deployed and in active use.

By the end of next year (2009) those numbers will have swelled to 53,000 labs serving some 52,000,000 students.


52 million students is very different from 52 million installations. They're all in labs which are (hopefully) used by many students each. The article is correct in drawing the only conclusion it can:

...there are at least 15 million active desktop Linux users out there.


Maybe they should have said that the number of acutal users could be much higher due to some of those installations being part of labs but that's not a conclusion the data lets them draw.

Reply Score: 2

RE[2]: Numeric discrepancy
by lemur2 on Mon 16th May 2011 04:42 UTC in reply to "RE: Numeric discrepancy"
lemur2 Member since:
2007-02-17

Maybe they should have said that the number of acutal users could be much higher due to some of those installations being part of labs but that's not a conclusion the data lets them draw.


The number of desktop Linux users is set to accelerate in BRIC countries and developing nations.

http://opendotdotdot.blogspot.com/2010/12/putin-orders-russian-move...

http://www.computerworld.com/s/article/9201699/China_OS_makers_part...

http://www.olpcnews.com/

I can't see that it matters much if some of these expanding numbers of users are using the same machine.

Reply Score: 4

Poor metrics
by wirespot on Mon 16th May 2011 07:29 UTC
wirespot
Member since:
2006-06-21

As much as I'd like to get a clear picture of Linux usage, the current methods of approximation are faulty. Unfortunately, there's simply no way to get any close-to-real data.

StatCounter tracks ~3 million sites out of over 100 million out there (Google estimate of 2 years ago). That's a very small sample. Also, it does not take into account the habits of Linux users and what sites they are more likely to visit. It simply picked some completely random sites and examined their traffic.

It's also rather easy to find data sources that make you strongly doubt the 15 million figure. For example, in 2010 Ubuntu alone was claiming 12 million desktops, and I find it hard to believe they alone account for 80% of the Linux desktops, given the relative popularity of PCLinuxOS, Debian, Fedora, OpenSUSE, Mint and Arch, to name just a few of the more widely used distros.

Reply Score: 10

RE: Poor metrics
by Lennie on Mon 16th May 2011 09:27 UTC in reply to "Poor metrics"
Lennie Member since:
2007-09-22

These are probably the best numbers you can get:

http://stats.wikimedia.org/archive/squid_reports/2011-03/SquidRepor...

You know what sites the users visited, it is all the WikiMedia sites, so mostly Wikipedia.

But Wikipedia gets a lot of visitors Alexa says they rank 8 worldwide: http://www.alexa.com/siteinfo/wikipedia.org

At the end of the day, it just says (this was March): Windows XP is at the top of the list, which is kind of sad. My guess is even Microsoft isn't to happy about that. It also makes webdevelopers sad, because it means IE8 will remain for a long time. As there is no IE9 for Windows XP.

Android has the largest Linux share, Ubuntu has the largest desktop share.

The share of Mac PPC users is still bigger than all the other Linux distributions and mobile devices.

Ubuntu has 41% of the Linux desktop share:

2.74% of Linux users, minus 1 for Android:

(100÷(2.74−1))×0.72=41.37931..whatever %


Here are some numbers per country:

http://stats.wikimedia.org/archive/squid_reports/2011-03/SquidRepor...

Edited 2011-05-16 09:35 UTC

Reply Score: 4

RE: Poor metrics
by WereCatf on Mon 16th May 2011 16:09 UTC in reply to "Poor metrics"
WereCatf Member since:
2006-02-15

It's also rather easy to find data sources that make you strongly doubt the 15 million figure. For example, in 2010 Ubuntu alone was claiming 12 million desktops, and I find it hard to believe they alone account for 80% of the Linux desktops, given the relative popularity of PCLinuxOS, Debian, Fedora, OpenSUSE, Mint and Arch, to name just a few of the more widely used distros.


I personally believe the 15 million figure is quite right, but I doubt Ubuntu's claim of 12 million desktops. It's likely a lot lower in reality.

Reply Score: 2

RE[2]: Poor metrics
by Morty on Mon 16th May 2011 17:47 UTC in reply to "RE: Poor metrics"
Morty Member since:
2005-07-06

I personally believe the 15 million figure is quite right, but I doubt Ubuntu's claim of 12 million desktops. It's likely a lot lower in reality.

Ubuntu claims always seem to be more hype than substance, so doubt those numbers may be right.

But I'd say that 15 million is on the low side, with the big distributions adding it all up. Something like 25% higher sounds more correct.

With OpenSUSE at about 7-8 million and Fedora around 4-6 million. Even with some recession the last year, it still has a size-able install base in South America and Asia. Mandriva should at least clock in at 2-3 million. They are always ignored when people pull up Linux user numbers. But you should always include two distributions that are huge in Asia, Red Flag and Turbolinux. They should at least add up to a few millions user, but I would not be surprised if they eclipse even the possible inflated number of Ubuntu.

Reply Score: 2

RE[2]: Poor metrics
by lemur2 on Tue 17th May 2011 05:04 UTC in reply to "RE: Poor metrics"
lemur2 Member since:
2007-02-17

I personally believe the 15 million figure is quite right, but I doubt Ubuntu's claim of 12 million desktops. It's likely a lot lower in reality.


If all of Linux desktop is a mere 15 million (despite there being 52 million Linux desktops for Brazillian students ???), and Ubuntu is indeed a lot lower than 12 million, then someone has got the utterly wrong idea and has set themselves an impossible goal:

http://www.techdrivein.com/2011/05/goal-is-200-million-ubuntu-users...

"Our goal is 200 million users of Ubuntu in 4 years", said Canonical founder Mark Shuttleworth.

So, I'm wondering if it is Mark Shuttleworth or WereCatf who might have a better idea of the current number of Ubuntu installations?

Edited 2011-05-17 05:07 UTC

Reply Score: 3

RE[3]: Poor metrics
by saynte on Tue 17th May 2011 06:16 UTC in reply to "RE[2]: Poor metrics"
saynte Member since:
2007-12-10

There are not 52 million Linux desktop machines in Brazil. I believe what you're referring to is the Brazilian schools being outfitted with Linux machines in the computer labs, allowing 52 million students to access the machines.

Even given that, the 52 million number is probably a bit optimistic even seen as a number of Linux users. 52 million is likely the total number of all school children in Brazil, so that's an incredibly inaccurate way to count.

Reply Score: 1

RE[4]: Poor metrics
by lemur2 on Tue 17th May 2011 10:12 UTC in reply to "RE[3]: Poor metrics"
lemur2 Member since:
2007-02-17

There are not 52 million Linux desktop machines in Brazil. I believe what you're referring to is the Brazilian schools being outfitted with Linux machines in the computer labs, allowing 52 million students to access the machines.

Even given that, the 52 million number is probably a bit optimistic even seen as a number of Linux users. 52 million is likely the total number of all school children in Brazil, so that's an incredibly inaccurate way to count.


There may not be 52 million Linux desktop machines in Brazil, but there are 52 million Linux desktops in Brazil. There are 52 million Linux desktop users in Brazil. More in fact, because the 52 million only counts the Brazilian students.

Reply Score: 3

RE[5]: Poor metrics
by saynte on Tue 17th May 2011 11:13 UTC in reply to "RE[4]: Poor metrics"
saynte Member since:
2007-12-10

I think the habit of counting all the people that could use the Linux installations is a bit ineffective to try to count users.

For example, to transpose the language we're using to another setting: all students in Brazil are baseball players because all schools have access to a baseball diamond. Ergo, Brazil has 52 million baseball players.

Reply Score: 1

Global figures
by twitterfire on Mon 16th May 2011 16:47 UTC
twitterfire
Member since:
2008-09-11
RE: Global figures
by jtfolden on Mon 16th May 2011 23:29 UTC in reply to "Global figures"
jtfolden Member since:
2005-08-12

Google trends aren't very telling, though...

Reply Score: 2

v Comment by westlake
by westlake on Tue 17th May 2011 01:13 UTC
Half of enterprise SW will be OSS
by lemur2 on Wed 18th May 2011 00:23 UTC
lemur2
Member since:
2007-02-17

A topic tangentially related to this thread is discussed here:

http://www.networkcomputing.com/virtualization/survey-half-of-enter...

Survey: Half Of Enterprise Software Will Be Open Source Within Five Years

"A survey being released today at an open source conference in San Francisco shows that 56 percent of respondents predict that more than half of the software purchases made by businesses and other enterprises over the next five years will be of open source software. The organizers of the Open Source Business Conference 2011 say that this is because customers have overcome their reluctance towards using open source, such as concerns about licenses, and are embracing its virtues, such as flexibility, lower cost and avoiding vendor lock-in."

Reply Score: 3