#+INCLUDE: ../org-templates/level-1.org
#+title: GNU/Linux users per 1000 windows users
#+date: <2020-12-22 Di>
#+LANGUAGE: en
TLDR: In 2020 the fraction of Linux users doubled compared to Windows
users. Compared to 2008 the Linux-user fraction rose by factor 5.
Every year people ask “is it finally the year of the Linux Desktop?”.
In 2020 the answer is clear: since 2008 almost every year was a year
of the Linux Desktop: Compared to Windows users there was reliable
superlinear growth:
#+begin_export html
<div class="sourcebox" markdown="1" >
<a href="gnu-linux-desktop-share.pdf"><img width="168" height="235" title="PDF" src="../assets/pdf-thumbnail.png" /></a><br />
<a href="gnu-linux-desktop-share.pdf">PDF</a> (drucken)
</div>
#+end_export
#+begin_src gnuplot :exports none :var netmarketshare=gnu-linux-per-kw statcounter=gnu-linux-per-kw-statcounter :file gnu-linux-desktop-share.png :exports results
reset
set title "GNU/Linux users per 1000 Windows users"
set key top left
set ylabel "GNU/Linux desktop share / GNU/kw
# set label 'systemd in Debian' at 2011,30
# set arrow from 2015,28 to 2015,18
set label 'LiMux reversed' at 2013,34
set arrow from 2018,32 to 2018,21
' set label 'covid' at 2019,14
' set arrow from 2020,16 to 2020,41
' set label 'homeschooling' at 2018,10
' set arrow from 2021,12 to 2021,20
set terminal png size 420,320
plot netmarketshare u 1:2 with lines lw 2 lt rgb "#1E90FF" title "netmarketshare", \
statcounter u 1:6 with lines lw 2 lt rgb "#1EFF90" title "statcounter"
#+end_src
#+attr_html: :alt A plot of GNU/Linux desktop share, plotted as number of Linux users per 1000 Windows users.
#+attr_html: :alt It shows two lines, one for each datasource: netwarketshare from 2008 to 2023 and statcounter from 2009 to 2024. Each is sampled every July.
#+attr_html: :alt Both start at 10, then go up to about 20 in 2018. THen netmarketshare jumps to 40 in 2020 while statcounter just keeps rising. In 2021 netmarketshare dips to about 15 again, but statcounter grows to 30.
#+attr_html: :alt Netmarketshare ends at 30 in 2023, while statcounter grows 42 in 2023 and then to 60 in 2024.
#+RESULTS:
[[file:gnu-linux-desktop-share.png]]
The source of this graph is data I collected over the past decade
every July from [[http://www.netmarketshare.com/os-market-share.aspx?qprid=9][OS market share]]. Except for the 2022 data which is
from October because I missed taking the data point in time. They do
browser detection, so the data is far from perfect, but at least they
have data we can use:
# TODO: replaced this first line due to a bug in org breaking the table extraction. Should be fixed in 9.5.3
# | 2008 | 8.6 | GNU/Linux users per 1000 Windows users (GNU/kw) |
#+name: gnu-linux-per-kw
| 2008 | 8.6 | GNU/kw |
| 2009 | 11 | GNU/kw |
| 2010 | 10 | GNU/kw |
| 2011 | 11 | GNU/kw |
| 2012 | 12 | GNU/kw |
| 2013 | 14 | GNU/kw |
| 2014 | 19 | GNU/kw |
| 2015 | 18 | GNU/kw |
| 2016 | 22 | GNU/kw |
| 2017 | 25 | GNU/kw |
| 2018 | 21 | GNU/kw |
| 2019 | 24 | GNU/kw |
| 2020 | 41 | GNU/kw |
| 2021 | 20 | GNU/kw |
| 2022 | 27 | GNU/kw |
| 2023 | 30 | GNU/kw |
As a second datasource, here’s this data as reported [[https://gs.statcounter.com/os-market-share/desktop/worldwide/#monthly-200901-202408][from statcounter]]
(postprocessed from their [[https://gs.statcounter.com/os-market-share/desktop/worldwide/chart.php?device=Desktop&device_hidden=desktop&statType_hidden=os_combined®ion_hidden=ww&granularity=monthly&statType=Operating%20System®ion=Worldwide&fromInt=200901&toInt=202506&fromMonthYear=2009-01&toMonthYear=2025-06&csv=1][csv export]] to be GNU/kw and filtered to just
july; and ignoring Chrome OS):
#+name: gnu-linux-per-kw-statcounter
| Date | Windows | Linux | Chrome OS | Unknown | GNU/kw |
|------+---------+-------+-----------+---------+--------|
| 2009 | 94.83 | 0.76 | 0 | 0.13 | 8 |
| 2010 | 93.29 | 0.77 | 0 | 0.08 | 8 |
| 2011 | 91.87 | 0.76 | 0 | 0.07 | 8 |
| 2012 | 88.85 | 0.83 | 0 | 0.06 | 9 |
| 2013 | 90.49 | 1.23 | 0.03 | 0.08 | 14 |
| 2014 | 89.25 | 1.34 | 0.19 | 0.63 | 15 |
| 2015 | 87.7 | 1.52 | 0.26 | 1.9 | 17 |
| 2016 | 83.33 | 1.54 | 0.35 | 5.16 | 18 |
| 2017 | 84.46 | 1.79 | 0.45 | 1.97 | 21 |
| 2018 | 82.88 | 1.71 | 0.5 | 2.37 | 21 |
| 2019 | 77.61 | 1.65 | 0.54 | 7.03 | 21 |
| 2020 | 77.74 | 1.85 | 0.67 | 2.67 | 24 |
| 2021 | 73.04 | 2.38 | 1.21 | 7.93 | 33 |
| 2022 | 75.21 | 2.76 | 1.5 | 6 | 37 |
| 2023 | 69.52 | 3.12 | 3.24 | 3.69 | 45 |
| 2024 | 72.1 | 4.44 | 1.41 | 7.13 | 62 |
#+TBLFM: $6='(round (/ (* 1000 $3) $2));N
/Thanks to [[https://www.gamingonlinux.com/2024/09/linux-smashes-another-market-share-record-for-august-2024-on-statcounter/][gamingonlinux]] and [[https://www.youtube.com/watch?v=qx3TvqUptLY][The Linux Experiment]] (video) for
referencing the statcounter data./
*GNU/Linux is growing.* (though there is a dip during the pandemic)
When will you migrate? Did you migrate already?
------
/This is based on an older German article I wrote in 2010 and updated regularly: [[http://www.draketo.de/licht/politik/leserbriefe/gnulinux-gewinnt-marktanteile][GNU/Linux gewinnt gegenüber Windows und MacOSX an Marktanteil]]/
* FAQ
** Could the increase be due to Windows users moving to mobile platforms?
Windows users did not go down enough, and MacOS did not rise along, to
warrant a 100% increase of GNU/kw. It’s at 87% while GNU/Linux in July
was only 3.57%. If this were due to an exodus of Windows-users, I would
expect MacOS to rise along.
The results are filtered for Desktop. When including mobile, Windows
did not have sufficiently high losses compared to mobile to explain
the Linux numbers differently than with an increase of Linux users.
# #+toc: headlines 2
#+setupfile: ../org-templates/level-1-politik.org
# Local Variables:
# org-confirm-babel-evaluate: nil
# org-export-default-language: en
# org-babel-noweb-wrap-start: "{{{"
# org-babel-noweb-wrap-end: "}}}"
# org-latex-images-centered: nil
# End:
# override template options:
#+options: toc:nil H:1 num:nil ^:nil