- Site menu:
- Latest Discussions
- Latest Newswire
- Archives
- Submit Story
- Story HOWTO
- Syndicate
- E-mail Alerts
The World According to Linux
It's time to realize just how inaccurate all of these web statistics are. They're another source of FUD against the Linux community. By implying that, despite all of the sales of Linux in the past year, that the number of Linux users has somehow stayed exactly the same, it scares would-be Linux users away who are afraid of problems with support and trust those big, bad proprietary corporations. |
|
|
Comments (15)
Maybe if we were to get stats from hits on distribution update servers.
tracyanne, not to be mean, but I am sure that you can blow large holes in that idea without even trying.
Also, I am a human and a linux user, not a metric.
Also, I am a human and a linux user, not a metric.
Quoting:I am a human and a linux user, not a metric
So am I, but just once I'd like like to have a definitive statistic that I can throw in the face of the corporate lie that Linux is less than consequential.
Some distros are quite desperate in this regard. But most of their efforts are thwarted by "privacy concerns" ;)
> Also, I am a human and a linux user, not a metric.
I'm not a metric. I'm an imperial.
I'm not a metric. I'm an imperial.
I registered with the Linux Counter project a while ago http://counter.li.org/ - I don't care if they know I use Slackware and live in Colorado.
Funny someone mentions the Linux-counter. There's a page where the number of people using Linux is estimated:
http://counter.li.org/estimates.php
Coincidentally (or maybe not), I'm the one who updated that page and made that 29M estimate (you can see LXer's migration list is mentioned ;) ). Before I made that estimate, I also thought about ways to make a better estimate of course. My conclusion was, that would be almost impossible, except if you go out on the street and ask a representative (meaning big very diverse and coming from diverse continents) population if they use Linux or not.
The distribution update servers are the best way to estimate probably. However, people can run their own local 'repositories' if their company has 100 computers running Linux. Then only one of those 100 machines would be counted in the distribution statistics. Also, people without internet will probably not update their machines. On the other hand, they can buy a CD/DVD with Linux, and copy this to give it to friends as many times as they want. So you should also involve the number of CD/DVD's Mandrake and (K)Ubuntu ship, and some governments make them available for free too. There are also hundreds of shops selling those CD/DVD's. When I / my parents had to pay per MB of downloaded data, it was cheaper to buy a Gentoo-CDROM for €10 than downloading it. Than there's a number of lucky people who receive Linux pre-installed with the computer they buy. This is especially true for Asia, but there's also the OLPC, the Asus Eee, the Ubuntu Dell lap/desktops, the Everex Linux boxes at WallMart, those at Carrefour and Tesco in Europe, the 'smaller' Linux shops, Zonbu and probably some others I missed.
If you add the number of people using the 'distribution update' servers to the number of people who bought a Linux CD/DVD or received it for free from Ubuntu, you'd probably have a lower limit. However, not everybody that receives a free Ubuntu CD ends up using it. And you cannot simply add the number of sold Dell's coming pre-installed with Ubuntu because those people will probably use the update servers too. However, others might not. The number of sold Eee's can be used neither, since some of them will be shipped with WinXP in the future. And Google is another difficult one: They use thousands of Linux-boxes, but how many Linux _users_ are they? And how would you count the people who only use a Linux box as a firewall / gateway in their home-network while all clients run WinXP? So this is all a big mess, where it's just as easy to count people twice as to not count them at all. Therefore, it's only possible to make a rough estimate.
I'd say between 50/60M these days, and at the same time I'd encourage you to come up with a better estimate, while at the same time knowing that's very difficult.
http://counter.li.org/estimates.php
Coincidentally (or maybe not), I'm the one who updated that page and made that 29M estimate (you can see LXer's migration list is mentioned ;) ). Before I made that estimate, I also thought about ways to make a better estimate of course. My conclusion was, that would be almost impossible, except if you go out on the street and ask a representative (meaning big very diverse and coming from diverse continents) population if they use Linux or not.
The distribution update servers are the best way to estimate probably. However, people can run their own local 'repositories' if their company has 100 computers running Linux. Then only one of those 100 machines would be counted in the distribution statistics. Also, people without internet will probably not update their machines. On the other hand, they can buy a CD/DVD with Linux, and copy this to give it to friends as many times as they want. So you should also involve the number of CD/DVD's Mandrake and (K)Ubuntu ship, and some governments make them available for free too. There are also hundreds of shops selling those CD/DVD's. When I / my parents had to pay per MB of downloaded data, it was cheaper to buy a Gentoo-CDROM for €10 than downloading it. Than there's a number of lucky people who receive Linux pre-installed with the computer they buy. This is especially true for Asia, but there's also the OLPC, the Asus Eee, the Ubuntu Dell lap/desktops, the Everex Linux boxes at WallMart, those at Carrefour and Tesco in Europe, the 'smaller' Linux shops, Zonbu and probably some others I missed.
If you add the number of people using the 'distribution update' servers to the number of people who bought a Linux CD/DVD or received it for free from Ubuntu, you'd probably have a lower limit. However, not everybody that receives a free Ubuntu CD ends up using it. And you cannot simply add the number of sold Dell's coming pre-installed with Ubuntu because those people will probably use the update servers too. However, others might not. The number of sold Eee's can be used neither, since some of them will be shipped with WinXP in the future. And Google is another difficult one: They use thousands of Linux-boxes, but how many Linux _users_ are they? And how would you count the people who only use a Linux box as a firewall / gateway in their home-network while all clients run WinXP? So this is all a big mess, where it's just as easy to count people twice as to not count them at all. Therefore, it's only possible to make a rough estimate.
I'd say between 50/60M these days, and at the same time I'd encourage you to come up with a better estimate, while at the same time knowing that's very difficult.
Quoting:Maybe if we were to get stats from hits on distribution update servers.
I recall a thread on LXer, which this issue was thoroughly discussed. May be one of LXer editors could dig it up. It was interesting and extensive about the various ways this could be done in fairly accurate ways.
My Toshiba came with their 'Express Media Player' which plays CDs and DVDs. It is running Linux and they put the GPL on the XP partition in the documentation as well as having a Linux documentation partition with all the source.
Then there are all those routers, set top boxes, cell phones running Linux.
Not many people are looking for a reason to give M$ more money, and they don't have to thanks to FOSS.
Then there are all those routers, set top boxes, cell phones running Linux.
Not many people are looking for a reason to give M$ more money, and they don't have to thanks to FOSS.
Wow, I have to say I am amazed here. Every last one of us derides M$ for their "dial home" check in, yet here are some of the same people just off handedly suggesting the same thing for Linux. Maybe not for the same purpose, but still as they say, it's the thought that counts.
Lets count em when they update: Accurate as a fart in a high wind, as hk pointed out. Debian popularity contest: That about equals M$ dial home, but with an opt out option. Registered Users: *cough* excuse me, isn't one of the things we have touted for so long the LACK of user registration?
Lets put this another way. Linux Advocate: "We need a way to get actual numbers" M$ Representative : "Here, let us show you how. We have done it for years"
Get the picture?
Lets count em when they update: Accurate as a fart in a high wind, as hk pointed out. Debian popularity contest: That about equals M$ dial home, but with an opt out option. Registered Users: *cough* excuse me, isn't one of the things we have touted for so long the LACK of user registration?
Lets put this another way. Linux Advocate: "We need a way to get actual numbers" M$ Representative : "Here, let us show you how. We have done it for years"
Get the picture?
> Debian popularity contest: That about equals M$ dial home, but with an opt out option.
Hardly "opt-out", much more "opt-in".
On the rest, I agree. We might just have to do what MacDonald's did and just say "Billions and Billions Served."
Hardly "opt-out", much more "opt-in".
On the rest, I agree. We might just have to do what MacDonald's did and just say "Billions and Billions Served."
Quoting:Every last one of us derides M$ for their "dial home" check in, yet here are some of the same people just off handedly suggesting the same thing for Linux.
I don't think I'm suggesting that.
Mandriva have a registration form at first time start up. It registers you with the Mandriva Forums, and enables Mandriva to collect some demographics. I've started filling this out for my new Linux users, in part so that they can have default access to the forums, and because I get to have the credit for being the person who put them on to Linux. If every Linux user registered we would know how many Linux usrs there are. If we added to that the number od desktops in corporate/business use, and hanging off LANS (assuming we can collect that information), we would have a very good idea of how many Linux desktops there are.
The other way of counting, that was suggested is the hits on update servers.
Every Linux distro "phones home" every time it checks for updates.
All the Distros would have to do is count the discrete IP addresses (this wouldn't count multiple set ups on a LAN, but it would go a long way towards giving us a head count. I personally don't have a problem being counted as a Linux user, and most people who get tetchy about giving out personal information to their OS Distributor give away far more personal information to far less savoury people/corporations simply by by having an email account.
Quoting: I've started filling this out for my new Linux users, in part so that they can have default access to the forums, and because I get to have the credit for being the person who put them on to Linux.
Only with their full knowledge and consent, right?
Quoting: I personally don't have a problem being counted as a Linux user, and most people who get tetchy about giving out personal information to their OS Distributor give away far more personal information to far less savoury people/corporations simply by by having an email account.
Your choice; any answer to this overall questions must involve others' giving their consent, each individually.
> Every Linux distro "phones home" every time it checks for updates.
Only if the various mirrors also keep count. I doubt the various Slackware mirrors do. It also doesn't solve the problem of those who update multiple machines from a single local server.
Only if the various mirrors also keep count. I doubt the various Slackware mirrors do. It also doesn't solve the problem of those who update multiple machines from a single local server.
Quoting:Only with their full knowledge and consent, right?
Yes, they are usually sitting beside me. [EDIT I to they]
Quoting:any answer to this overall questions must involve others' giving their consent, each individually.
Indeed. I would like to reiterate.
Quoting:most people who get tetchy about giving out personal information to their OS Distributor give away far more personal information to far less savoury people/corporations simply by by having an email account.
Post a comment
A free account is needed to post. Write your comment here first: your text is kept in this browser while you sign up or log in, and is waiting for you when you come back.
Your message:Markdown: **bold**, *italic*, [text](https://link), - lists, > quotes, $x^2$ for math. Formatting help
Sign up and post (it is free) or Log in and post