Windows vs. Linux Market Share: Where are the facts?

[DrJonesAC2]

I am doing a research paper for my Economics on the effect of linux on windows market share. I have been google-ing for days and have not found any actual, realiable, stone-cold statistics. Does anyone know where I can get this information? All I find is propaganda for both sides. IT is very frustrating. I appreciate any help offered.

[Choozo]

I don't think you will ever find any hard and reliable facts on this subject.

1. Microsoft count every box shipped (also those still in store shelves) and every PC with Win pre-installed. Additionally, when people upgrade from one version to a newer (e.g. Win98 -> WinXP), the previous install do not get deducted from their numbers.

2. Linux distributions are freely distributable, so counting downloads and boxed sets sold will not get you any reliable numbers. Additionally, people bying a PC with Win pre-loaded may delete Win and install Linux instead - and this removed copy of Win won't be deducted from Microsofts count either.

[DrJonesAC2]

But I there has to be a source (skewed yes) that is an authority with actual numbers.

[El_Cu_Guy]

Microsoft count every box shipped (also those still in store shelves) and every PC with Win pre-installed.

Who cares about boxes? Microsoft counts licenses. Audits are another way of keeping track.

Tracking the number of downloads is easy. However, this isn't 100% accurate either. It doesn't factor in the number of corrupt downloads (and re-downloads) and considers ever downloaded copy to have been installed.

So you see there are two side to every argument about market share. Niether is 100% accurate.

LinuxWorld (http://www.linuxworld.com/2003/0314.petreley.html) has an article (March 14, 2003) titled Debunking the Linux-Windows Market Share Myth.

[DrJonesAC2]

IT seems that choozo may be right...There may not be any actual factual data out there. But where does cnet and the otheres get thier information from when they announce that Linux has gained such and such marketshare etc....
At this point in the game I really dont care if the information is 100% correct I just need a fairly reliable source that gives the big picture ( a bar graph would be nice)

[mrBen]

Here are the facts:

1. There _are_ more windows boxes than Linux in the world.

2. Linux _is_ catching up.

That's about all we really know ;)

[lallepot]

To answer the question on where all the news source get the hard fact...

My experience with news agencies is that the facts or story never makes the news. The definition of "hard facts" must be backwards. "Hard facts" is what the news brings - as long as its a good story. Credibillity of the source is a matter for historians and not the news.

The % of use of OS is properly determined by sales figures - which ofcause is not the same as the % of use.

To get an estimate of the use of OS' I would check out the browser statistics - even they only include % of OS used for http browsing.

Lallepot

[joelc]

Two things.

1. Market share (% sales) vs Actuall Use (% of active machines running linux/windows).


2. There is no way to determine actual use. If you need a little filler in you paper you could devote a small paragraph to the problems involved in that alone. The closest we're going to get is the statics based on web browsing. There are places (though I don't remember any names of the top of my head) that keep track of that keep track of that kind of statisitc on the web (what percent is IE, what percent netscape, what percent opera, and what operating system are they using), but even then some browsers let you pretend to be somebody else (so that if you come to sights that treat IE 'better' you can pretend to be IE.

[nerrollus]

Originally posted by El_Cu_Guy
Who cares about boxes? Microsoft counts licenses. Audits are another way of keeping track.

Tracking the number of downloads is easy. However, this isn't 100% accurate either. It doesn't factor in the number of corrupt downloads (and re-downloads) and considers ever downloaded copy to have been installed.

So you see there are two side to every argument about market share. Niether is 100% accurate.

LinuxWorld (http://www.linuxworld.com/2003/0314.petreley.html) has an article (March 14, 2003) titled Debunking the Linux-Windows Market Share Myth.


Counting licenses is the way to count MS. The company I would for only has like 5 windows NT / w2k install disks, but we actually have nearly 400 working licenses for just our building. This is a world wide company, so I'm guessing that everyone is the same way. =)

[CMonster]

It's all relative to how you count -isn't it?

www.google.com "linux market share" now scroll and click and see if you can gather a few kernels of wheat amidst the all the weeds


Also try a search at www.idc.com -be prepared to enter very specific search criteria like "linux worldwide market share" or "linux north american market share"

Point of interest:

E-commerce solutions on Linux comprise 23 per cent of today's Linux revenue in the marketplace and is expected to grow to 40 per cent by 2004. As part of its effort to tap the Linux space, IBM plans to have a whole family on Linux: Tivoli support for server management and national language support in the future, Kendra said.

This year, IBM will focus on front-end tools for usability, implementation, installation and customisation aimed at providing customers with an end-to-end solution supported by integration and existing systems application.

http://www-3.ibm.com/software/webservers/p010307.html

[Darksamurai]

Might take into account the fact that Wal-Mart has started making PCs with Lindows. That might give you somewhat of a base if you hit lindows.com, or walmart.com...

It wouldn't be definitive... but it might give you enough info for your paper!

[teeitup]

I manage a software test lab.

About a month ago I was surveyed about server use.

It was very enlightning. The survey didn't care what I was using only what I purchased brand new in the last 18 months.

I had no way to indiacate that 10 Proliant servers were now running Lunix. Nt was deleted and the servers now handle software version management. deployment, file storage, bug tracking and isolation from the production network.

We've increased the number of servers.
We've reduced the Microsoft licence costs.

We've also reduced the Sun servers and licenses and the survey had no way to reflect any of this.

Be careful of surveys in the IT world. There is usually a result in mind when the survey is drafted and paid for.

Good Luck,