Removing market share FUD the linux way...

Forum: LinuxTotal Replies: 18
Author Content
knoba

Jun 28, 2009
7:38 AM EDT
I'm fairly certain that I'm not the first person to ask this; but wouldn't a small program, portable for every Linux distribution, be able to register with an on-line database about how many users actually run Linux on-line? Debian already has an optional "Popularity contest" program that sends out information about what software packages are installed on the users' system. If the poll was anonymous (and voluntary) I wouldn't mind sending out that information as a Linux/GNU/FOSS advocate. If true Linux market shares were known it could help the entire community/movement in many ways.

If anyone knows of such a program; could you let me know please.
jdixon

Jun 28, 2009
10:27 AM EDT
You don't need a program. Simply register with the Linux counter project: http://counter.li.org/

That's as cross platform as you can get.
Sander_Marechal

Jun 28, 2009
10:45 AM EDT
You already answered your own question. "optional". To make any kind of reasonably accurate guess that's not good enough.

Except for doing a full, complete market study there's no way to figure out how big Linux is. Anything else will only give you a lower bound (which might still be interesting though).
knoba

Jun 28, 2009
6:41 PM EDT
@ jdixon - It doesn't seem to be a good way to estimate figures though. I registered with them years ago, but lost my membership number & email address. I could sign up again, but it makes me wonder that the system is open to abuse. It just doesn't seem to hit the spot.

@ Sander_Marechal - I wouldn't have thought it would be too far off the mark. Most Linux / GNU users are happy to contribute to the community. A real-time, anonymous, optional, on-line counter could help considerably (rather than relying on browser ID's, counters, etc.).

I thought that; (in-keeping with the Unix philosophy - programs that do one thing and do it well) Linux would have and benefit from such a program. Anyone know of such a thing?
Sander_Marechal

Jun 28, 2009
7:16 PM EDT
Quoting:Most Linux / GNU users are happy to contribute to the community.


Not in this case. The vast majority of Linux runs in corporate environments on servers. These will most likely not be counted by such a program.
hkwint

Jun 29, 2009
4:52 PM EDT
Quoting:Anyone know of such a thing?


Sure; it exists and lives here: http://counter.li.org/scripts/

Someone should convince Linux-distro's to include it by default (opt-in during installation time I guess) however. It's one more step in installation however, and one more question nagging the user. That's the reason why it's not in the distro's I guess.
helios

Jun 29, 2009
7:10 PM EDT
Scenerio:

I leave a HeliOS Solutions Customer, customer clicks on interesting icon which in turn pulls up a gui or text file, explaining the script hkwint offered.

Here is what NEW Linux User will see. Emphasis mine:

* Save machine-update to a file named "machine-update" in any convenient place * From a terminal window, change directory to that place. * Change it to be executable (chmod +x machine-update) * Run "./machine-update -i" to enter information * Run "./machine-update -m" to send mail * Run "./machine-update -c" to add it to your crontab file

nuff said?
jdixon

Jun 29, 2009
10:04 PM EDT
> Someone should convince Linux-distro's to include it by default

Slackware tells you about the Linux Counter in the welcoming email, but it expects you to be able to handle the registration yourself. :)
tuxchick

Jun 29, 2009
10:45 PM EDT
So if Slackware invites you to lunch, you should expect to bring and cook your own food? ;)
tuxchick

Jun 29, 2009
10:51 PM EDT
BTW, all of this agony over estimating Linux market share is overblown. On purpose, I'm sure, by the lackeys of Sauron. There are proven methods of collecting accurate data via surveys. Make them large samples of a number of different demographics, break it down into desktop, server, embedded, SuperLinuxComputingCluster of the universe, what have you. Throw in some self-selected surveys to contrast and compare. If you can't come within a couple of percentage points of good reliable numbers then you're doing it wrong. I bet money that Microsoft, IBM, Oracle, and other big companies have done these sorts of surveys numerous times.

The sticking point, of course, is funding good surveys. But it's really not a difficult problem, just an expensive one.
jdixon

Jun 29, 2009
10:54 PM EDT
> So if Slackware invites you to lunch, you should expect to bring and cook your own food? ;)

Bring, no. Cook, yes. Only you can know how you really want your food prepared. :)
gus3

Jun 29, 2009
10:55 PM EDT
Quoting:So if Slackware invites you to lunch, you should expect to bring and cook your own food? ;)
Only after I contact the National Enquirer.

As the saying goes, you can lead a horse to water, but if you can get him to do the back-stroke, then you've got something!
jdixon

Jun 29, 2009
10:59 PM EDT
> ...bet money that Microsoft, IBM, Oracle, and other big companies have done these sorts of surveys numerous times.

Microsoft doesn't need to. They get the sales figures from the major vendors, and they know how many licenses of Windows they get paid for. They can figure 90% of the difference is Linux. Then they count how many sales of Windows they get from places like NewEgg and work the numbers in reverse, adding in a fudge factor to allow for a higher percentage of do it yourself machines being Linux installs,
tracyanne

Jun 30, 2009
12:07 AM EDT
Quoting:As the saying goes, you can lead a horse to water.......


Or as Our late Premier Jo Bjelke Petersen would say "You can lead a horse to drink, but you can't make it water."
Sander_Marechal

Jun 30, 2009
3:00 AM EDT
Quoting:Microsoft doesn't need to. They get the sales figures from the major vendors


Sales figures are useless. In order to accurately get Linux market share numbers you need to do an actual market study. That is, get out there and interview/survey people. And great care must be taken that the sample is representative.
bigg

Jun 30, 2009
6:21 AM EDT
You have to first define what is meant by market share. Share of what? It's not straightforward.

In my house there are four laptops of various ages and three desktops. All came with Windows, and are set to dual boot several Linux distros. Windows market share might be 100%, Linux market share might be 100%.

If you are a hardware vendor thinking about supporting Linux, then what you need to know is that 100% of my dollars go to the purchase of hardware that has good Linux support.

A small business with 46 Windows machines and four Linux machines would be 92% Windows if you are talking about hard drives. However, when choosing to buy hardware, they're likely to go with hardware that is compatible with both Windows and Linux, when possible. I've seen something similar happen where I work. The appropriate unit is not "hard drives with Linux installed".
Steven_Rosenber

Jun 30, 2009
4:30 PM EDT
It really doesn't matter how many geeky types wipe Windows or Mac OS and install Linux (or BSD ... or OpenSolaris). What matters is preloads. That's quantifiable, it hits a different audience, and until FOSS gets more preloads, the whole thing is academic at best.
gus3

Jun 30, 2009
5:17 PM EDT
I partly agree The catch is how to show that there's a need for the pre-load?

Think of it this way: Linux has a certain percentage of the installations. Microsoft has a certain percentage of OS licenses sold.

If we restrict the discussion to PC's only (sorry Mac... well, not really), then the sum of these two percentages is probably greater than 100%, due to the overlap incurred by the "Windows tax".

I expect Neelie Kroes' henchpeople will resolve that long before the USA does.
number6x

Jun 30, 2009
5:53 PM EDT
gus3

Overlap due to dual boots as well.

You cannot post until you login.