I presume you have heard the term “circular definition,†Robert? If not, look it up.” A poll of ~1K randomly-selected people may well tell the story better than selecting 100K people from Winnipeg.ĭrLoser also wrote, ” If you add up random 1s and 0s you get a normal distribution. If you don’t shake you will never hit the bull. If you randomly fire while shaking the barrel a bit, you may hit the bull. overcounting A and undercounting B, a random re-sampling can in part remove the bias. There’s a lot of room to grow and with M$ being unable to count the opportunities continue.ĭrLoser wrote, ” Even random numbers would be better than no weighting…I suspect that we can both agree that you mis-spoke on that point.” The baseline has doubled in 5 years despite a lot of competition from Android/Linux. Still GNU/Linux is not dead and will rise again. That could be the result of M$ killing the netbook and along with it a bunch of new GNU/Linux PCs consumers could use or StatCounter rejigging their stats or my weights for ~2013 not being valid back then. This data is for all operating systems, not just desktop.Ī surprising result from my calculations is that GNU/Linux flattened around 2010/2011. That’s almost exactly what StatCounter puts out for their “unweighted” number of 1.17%. Now, all that remains is to add up the weighted sum and divide by that number…Īnd the answer is 1.18% for GNU/Linux’ global share of page-views. MariaDB > select sum(pop) as Total_Internet_Population from countries | Field | Type | Null | Key | Default | Extra |
![linux logtail norwap linux logtail norwap](https://i.ytimg.com/vi/nmljOGQOz7c/maxresdefault.jpg)
It’s surprising how many are using the Internet these days… People can object that the numbers are not accurate/up to date, but even random numbers would be better than no weighting… So, the number we should get will relate to the share of Internet-connected people who use GNU/Linux, more or less. Wikipedia has done the maths… There’s also, Internet World Stats Clearly, it’s Internet-connected people who make those page-views.
![linux logtail norwap linux logtail norwap](https://fullstackwebstudio.com/locations/wp-content/uploads/sites/18/2012/12/linux-a-login.jpg)
The weighting I chose is the population X percentage of “Internet-connected people” numbers from ITU.
LINUX LOGTAIL NORWAP FULL
Such dramatic growth should be shown in the best light, not thrown to the bottom of a trash heap full of M$’s saturated markets. Not knowing what the biases are is no excuse for not correcting the obvious one, that some regions are over or under-counted. I’m quite satisfied that the shape of this curve tells us something about the rate of adoption of GNU/Linux in Norway but I’m still skeptical about the magnitude and the merits of just throwing the number into an unweighted sum.
![linux logtail norwap linux logtail norwap](https://sc02.alicdn.com/kf/HTB1YuF6gpOWBuNjy0Fiq6xFxVXak/221590360/HTB1YuF6gpOWBuNjy0Fiq6xFxVXak.jpg)
Well, after a bug was fixed at StatCounter (no September data for any year could be downloaded before the end of September…) I’ve a database full of GNU/Linux shares for the world.