Google, one of the most visited sites worldwide, posted their OS results for April 2003. In the No1 spot, Windows 98 is steadily losing 1% every month to the always rising Windows XP, while MacOS is down to 3% from the 4% of the previous month. Linux is steady at around 1%. The “Other” OSes are also steady at 4% (note that the “other” section also includes other Microsoft OSes, like Windows ME).
The Top 8 Operating Systems According to Google
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Wow! Linux is an incredible OS with tons and tons of applications, infact it is becoming increasingly difficult to find any software which you can only run on Windows, and this is only with a 1% market share!!! Can you imagine what Linux would be like when it hits 2%! Linux may have 1% market share, but it’s got 50% of the developers! 😉
I was hoping that Linux would at least have pulled ahead of Win 95 by now. Tux still has a long way to go.
Actually, Linux has less than 1% market share currently. But it is rounded up by Google there (don’t forget that Google is using Linux, they would want to show it up
, plus Linux users only use Google pretty much. Other OSes are coming with netscape which defaults to Netscape search engine (and people get accustomed to it).
If you read some previous OSNews items reporting on IDG and 1-2 other big statistical companies, they all put Linux between 0.30% and 0.45%. That was a few months ago, today Linux might be around 0.55%, plus all these Linux servers that do not expose themselves to statistics easily, so Linux today might be around 0.70 and 0.80% overall.
Be careful with statistics. Linux’s market share is hard if not impossible to measure correctly, due to the fact that you can download it freely. I think (but that’s just speculation) that its market share is somewhere between 1 and 5%.
>I was hoping that Linux would at least have pulled ahead of Win 95 by now
No, certainly not. Windows 95 has still quite a large crowd (unfortunately, as I find Win95 a very unstable OS after you install third party drivers and applications), even bigger than Windows ME.
>Linux’s market share is hard if not impossible to measure correctly, due to the fact that you can download it freely
This is not true in the Google statistics. You see, Google is analyzing what people actually USE. Other statistic companies, might get into the trap your are expressing, because they might count sales instead of usage. But Google does not have that problem, therefore their stats are close to the actual numbers: around 0.80%-1% overall (including “faceless” servers)
>Linux’s market share is hard if not impossible to measure correctly, due to the fact that you can download it freely
Yeah and because us true OSS geeks use dmoz.org
Is there a place we can view statistics on osnews.com?
Not true; linux has _way_ over 1%. However, this really only counts for the desktop market, as no one really browses from their servers.
I don’t know about other people, but on my linux desktops I set the user agent to internet explorer, to get by lame sites that “only function with IE”.
Also, how does one accurately count web browsers? Per-IP? All but a few internet users (read home users, not servers/companies/etc) have dynamic IPs. If I go on today with my IE client and tomarrow my IP changes, any linux user that gets my previous IP won’t count.
Or, if google just counts total browser hits, then I’d assume that desktop users view google a lot more than the occasional sysadmin using lynx on a server to quickly look something up.
>Is there a place we can view statistics on osnews.com?
Yes. This is an external source for our site:
http://66.181.171.71/2/42699/7/ (monthly impressions: web page views per month (not hits))
General page: http://66.181.171.71/2/42699/0/?date=0 and here are the OSNews OS stats: http://66.181.171.71/2/42699/400/?sub_page=9&date=0
i propose that what investors are more interested in is not the _numbers_ using linux, but how much they spend, as a proportion of all computer spenders.
i look at my high-street newsagents. i see about 10 Windows PC magazines… about 2 Mac magazines… and about 2, sometimes 3, linux magazines. people buy these. advertisers advertise in these.
this weighted measure of say 10% is more reflective of active, rather than passive, users… people who upgrade.. people who tweak.. people who buy second hard drives… and i think companies should look at this active 10% and start writing drivers….
t
Desktop usage tends to “explode” with certain developments. Obviously the Linux desktop is pretty far behind Windows on many matters that count.
Not only are Gnu desktops hard to config and either look geeky or weird; but MSFT also works hard to preempt computer manufacturers from investing in Linux, by cutting down on demand. That’s why they can’t “lose” to Linux, because it would spur investment and accelerate Linux improvement.
it seems like the statistics is correct for users. but not for server deployments, since most servers don’t just ‘google by themselves’
.
another nice thing i can see from these stats, is that mozilla based browsers are about 5% (measuring the graph with a virtual ruler).
next thing google should figure out how to do, is to measure OpenOffice usage
I noticed that there where no reported Safari hits, I guess they are getting rolled in with the Mozilla ones.
http://sourceforge.net/softwaremap/trove_list.php?form_cat=199
The number of open source/free applications that run on my platform is about 15,000 more than those that run on Windows.
So yes, linux may have 1% of the users, but like I stated before, we got 50% of the developers, and that’s all that matters to me.
> I guess they are getting rolled in with the Mozilla ones
Yes.
Google Zeitgeist doesn’t say that 95% of the one word Google queries come from Windows users…
Or that Mac users are the source of 97% of the Jennifer queries.
It would be very interesting to see platform matched up with more usage information.
Which of course Google supplies to the USGOVT but not to the public.
Most interestimg to me is the inevitable replacement of Windows 98 by XP. And I’m suprised that many people still use Win 95…
Goes to show that most people will use whatever OS came installed on their computer as long as they keep that machine. It also says something about the number of “obsolete” computers quite capable of simple tasks such as web browsing.
-Bob
Good job that as an oldtimer I haven’t a bl..dy clue about all the search requests that google say THE people out there are interested in. Clearly from those requests 99.x% of PC users are also clueless about technology and therefore their preferences are utterly irrelevant to me.
The fact that I am using a patched Net+ve to look anon ie as Windows like as possible doesn’t bother me either. If you really are into the best technology, you are going to be less than 1% of the market.
Now if the stats could be honed down to the most power hungry user with high edu backgrounds, that would be much more interesting, probably this here sites stats would be valid enough.
The fact that there are so many Linux, Oracle, MS C++, Java, etc mags says alot more to me. I still remember NeXT had a MacWorld mag clone at one time. I don’t think the 99% ers buy any comp mags at all (games yes).
Something that wasn’t mentioned yet (I think
) is that most GNU/Linux users certainly dual boot and many probably still spend most of their time in Windows, simply because they can do more there (playing games, using their latest pirated graphics applications, etc).
So the actual share of people having Linux installed (or occasionally trying it) is certainly higher. But then again, Google is obviously biased considering that this is where I end up when typing a simple keyword in any of the free browsers.
It is also very interesting to see that many people still use Windows 98. I think those are the people which RedHat and Co should target first, because they might prefer to give a free alternative a chance (even if dual booting) instead of paying much money for Windows XP or a new computer (well, like me).
Comparing Windows 98 and RedHat 9, RedHat wins hands down in most areas (stability, security, visual appearance, functionality and usability) and just needs more work in simplicity, performance and overall polish IMO. Then it’s all about the applications.
Is that possible?
Eugenia, safari is based of konquerer, not mozilla.
Michael from pacbell.net wrote, presumably tongue in cheek: “Mac users are the source of 97% of the Jennifer queries”
What’s a “Jennifer query”? Never heard that term before. A quick search of google (using Lynx/Solaris9 for those keeping track) didn’t reveal anything.
Yours truly,
Jeffrey Boulier
>Eugenia, safari is based of konquerer, not mozilla.
And you think I don’t know that?
We are talking about how Safari is looking at the stats. And older stat engines are picking it up as Mozilla.
Mm, and IE is based on Mosaic. That doesn’t stop it from pretending to be Netscape.
As many pointed, statistics is something you need to spend a long thoughtful to really understand what’s going on. There are many reasons why windows 9x (and now XP) is so ahead in this kind of survey.
* Windows came installed in the vast majority of computers (and frankly, most of the users will keep using what came, switching only if something happen – big upgrade, severe system error, …);
* For a long time (a bit less now) the “online experience” was more “pleasant” in windows xx (better fonts and easy setup). I used to work in Linux and browse in windows xx environments (what I really still do);
* Many sites needed an ie browser to be displayed properly;
* Is still easier to setup a home computer with windows xx, and drivers are by far readily to find (they come with hardware installation cd-rom and are well distributed in the internet);
* A lot of pressure from microsoft on hardware manufactorers/integrators and software-houses;
* Better productity tools from MS (Office), Corel, Adobe and lots of games and educational programs (I know “better” here is very questionable by many reasons) because these are softwares that people are accustomed to use and will not switch easily;
The last point is strong, and changing (slowly) now (anyway, people don’t use OS because OS but because applications).
What I really want to know (what is not answered by this kind of survey) is how the things are changing in business environments, where many started to discover that isn’t needed “office assistants” to write a letter, verify mails or do some simple calculations in one spreadsheet, or that in many situations is straightforward to employ a *nix server. This is where microsoft started to be kicked and why they are so worried.
“Something that wasn’t mentioned yet (I think
) is that most GNU/Linux users certainly dual boot and many probably still spend most of their time in Windows, simply because they can do more there (playing games, using their latest pirated graphics applications, etc).”
Now THAT is funny!
Kinda OT, but funny:
The statistics spike in March. Is that when OSNews was almost slashdotted because of that Desktop Environment Comparison? Hahah!
“”Something that wasn’t mentioned yet (I think
) is that most GNU/Linux users certainly dual boot and many probably still spend most of their time in Windows, simply because they can do more there (playing games, using their latest pirated graphics applications, etc).”
Now THAT is funny! ”
Once someone tries Linux why would he or she want to go back?
Especially with this http://www.kde-look.org/content/pics/6225-1.png penguin popping up whenever he or she tries to log out of KDE. (Konqueror just wasn’t authorative enough.)
Thanks Eugenia for the OSNews statistics.
It very interestingly reveals that even if Linux/X11 is at around 35% of your user base, Mozilla is already used by more than half of your visitors.
That reveals the trend among software developers to gradually move over to open source. Eventually this will have an effect on the overall user base.
>Mozilla is already used by more than half of your visitors.
As I explained above, this stat engine doesn’t pick up all new browsers. If a browser is pretending to be mozilla (e.g. new opera versions, Safari) these will show up as Mozilla. That doesn’t mean that it is actually Mozilla, it could be anything. But indeed, 8 months ago, IE was the No1 browser of the people who were visiting OSnews, today is Mozilla, so there is a change indeed.
>Is that when OSNews was almost slashdotted because of that Desktop Environment Comparison? Hahah!
Where do you see the joke?
OSNews was Slashdotted 4 times on March and about 3 times on April IIRC. We usually get slashdotted as many times _every_ month, since September 2001.
March and April were big because we had literally _zillions_ of longer articles. Check our archives and you will see that we carried more than 30 longer “original” articles on March/April, including some world exclusive reviews. This is what made these two months big. May, will be more normal as I am currently taking a break and I only post newsbits. This month, I will only post 1 or 2 longer articles (in fact, the only review of mine for this month is going to be posted tomorrow), the rest will only be newsbits small articles like this one.
I have Konqueror set to identify itself as IE 5.5 on W2000. Some web pages work better that way. A few do not work at all without it. I don’t know if Google counts that as a Windows user.
Another place to get statistics is Netcraft, for example
http://uptime.netcraft.com/up/graph/?host=www.osnews.com
If you consider that there are hundreds of millions of computer users, then 1% is a signficant number. That single percentage point represents millions of people. Also consider that this 1% largely represents some very important catagories of computer users. It represents scientific and business workstations. It represents university CS machines. It represents computer-enthusiast desktop machines. All of these catagories consist of people that are active, and often avid computer users, and thus represent a market segment that has influence in a disproportionate amount relative to its size.
Linux has an important existing user base. Linux (and UNIX in general) software, in fact, excels in exactly the areas that are most important to this userbase — workstation applications, scientific applications, development tools, etc. I guess my point with all this rambling is that Linux developers, in their efforts to make Linux more palatable for a wider variety of users, cannot afford to alientate this user-base. It’s like politics. A republican candidate can go ahead and try to cour minorities and swing voters, but he can never afford to neglect staunch conservatives, because at the end of the day, those people are his base — the people who would vote for him anway. Thus, any major changes in Linux software will first and formost have to be palatable to existing Linux users. Pleasing your existing customers is a better idea than pissing them off in favor of new customers you may or may not get.
“If you consider that there are hundreds of millions of computer users, then 1% is a signficant number.”
Well, Google is just one search engine. Last I heard, they had about 50% (roughly) of the search engine market. Maybe there’s about 400 million Internet users worldwide (Or is that too much?). So 1% for Linux would amount to maybe 2 million Linux users using Google, assuming that Linux users and Windows users etc all use the same search engine (I’m not sure if that’s a true assumption either).
“That single percentage point represents millions of people.”
Only a few million, if that, I think.
The Windows figures totally swamp the Linux figures though.
“Also consider that this 1% largely represents some very important catagories of computer users. It represents scientific and business workstations. It represents university CS machines. It represents computer-enthusiast desktop machines. All of these catagories consist of people that are active, and often avid computer users, and thus represent a market segment that has influence in a disproportionate amount relative to its size.”
Yep. Linux users are generally particularly important people, more so than Windows and Mac users.
If you look at the previous months Apple’s share of the pie is:
March: 4%
February: 3%
January: 4%
I don’t think their share is going down.
I’m one of the Linux user who switched back to Windows for personnal usage:
– Mozilla on Windows is quite good, so I don’t need anymore to use IE/Outlook that I dislike (popup, viruses,etc)
– WindowsXP/2k are quite stable, stable enough for me.
My personnal usage is 50% web browsing 50% games, so if I keep Linux I would have to keep running two OS, while with Moz I can do all in Windows without trouble.
I really just think it shows that most people use Windows, and I’d bet that a fair amount of people who use Linux use Windows as their primary desktop. One example of where this seems to often be the case is when stats are run against some of the Linux-related newsgroups. Lots of people use Outlook who are apparently also Linux users.
I imagine that it is very true that most people just use what’s on their computer, and that most people aren’t very interested in their operating system. We live in a time where any major OS can do all of the tasks that a typical user needs; you can get e-mail, browse the web, and write word processing document on any computer. Most people don’t build their own PCs, and most that are bought come with Windows. Most people probably can’t find a reason to change.
I just don’t seem to think that market share or total usage is really all that important. I know there’s a big desire to make Linux more mainstream. If I was a gamer and needed those kinds of software, I might be one of that group myself, and that’s fair enough. But today I can get just about everything done in Linux other than video editing, that I’m having some issues with. I browse the web in Linux, do all of my IRC and e-mail in Linux, etc.
I can’t find any reason to recommend that most people I know should switch to Linux or another OS from Windows. Most don’t do much with their computers, and what they do spend time doing can be done about equally as well on any OS. I’ve recommended Linux (and recently FreeBSD) to people who are power users. Not necessarily developers or anything, but people who take computers a little more seriously.
I can see how more widespread adoption would be important to shareholders in Linux-related companies. But I personally find it hard to believe that any OS will ever die because of market share. Linux started with very few people, and I imagine it would survive and continue to develop with very few people. This may be the case with FreeBSD today.
I guess it would be nice to have some more people in the fold so that companies would put more and more money into the development of the free tools and applications all of we free OS users rely on. But I’m hardly disappointed with the way Linux is rolling along now. I read frequently how it’s not ready for the desktop, but I haven’t noticed this, for myself, and have now been using it as my desktop for about a year. All OSes could be better but I don’t think Linux’s readiness for the desktop is all that great an issue, frankly. It’s that people get Windows “for free” pre-installed, get used to it, and don’t need anything else.
Most non IE on windows browsers pretend to be that anyway so those problematic websites actualy have a chance at working.
Since people ask regularly for this site ‘s statistics, why not add a page with those links?
Just because a lot of software on Sourceforge are for other platforms, doesn’t mean that all developers use Sourceforge.
“This is not true in the Google statistics. You see, Google is analyzing what people actually USE.”
And how do they measure that? Anyone using a minority browser or OS is foreced to spoof as IE on Windows most of the time.
IMO finding out what share of the actual man hours of computer use each OS has is very difficult indeed.
It is stunning to see these stats when, like pretty much all of us, are using and trying different OSes all the time. I forget what most people are doing. Windows 95 – ack! 🙂
It would make sense that Windows would take the top spots on Google, since most Windows users haven’t figured out how to use Bookmarks/Favorites yet. They HAVE to Google just to get around.
How do they determine which browser you’re using? From the user agent string it returns?
If so, the results are guaranteed to be skewed heavily towards IE/Windows. This is because many non-mainstream browsers are set by default to report themselves as IE, otherwise some sites won’t let them in. A self-fulfilling prophecy, really…
I’d be willing to wager that people like the guys at ILM of Sandia National Labs, are more influential than grandma at home. The Linux user base has disproportionately more of the former than the later, which makes the Linux user base more influential than its size would suggest. Heck, even your Joe-average Linux user is generally a comuter enthusiast, which means that he/she spends more money and time on computers, and thus is more important to tech companies than an average Windows user, who is much less likely to be a computer enthusiast. It’s very similar to the gamer’s market. Gamers spend a huge amount of money on their computers, and as a result, gamers command a great deal more influence than their physical market size would suggest.
Opera 7.11 spoofs as IE by default, yet some sites still detect it as a ‘foreign browser’. So I guess there is an Opera tag somewhere in the referrer after all, to add some validity to these figures.
//So yes, linux may have 1% of the users, but like I stated before, we got 50% of the developers, and that’s all that matters to me.
//
Too bad only 10% of them can code worth a fart.
It would be interesting to see those stats broken down by time of day. I wonder how many are googling from work and therefore have no choice of platform.
Too bad only 10% of them can code worth a fart.
Are you referring to commercial software or free software? Can you back this up with examples? If you compare comercial software for linux vs commercial software for Windows, I don’t see any difference in quality standards. And comparing free software with Windows shareware, I would say that the Linux stuff is higher quality over all. Comparing free software with commercial software I don’t think is a fair comparison, but even in this case, I can site many examples of where free software is of higher quality than its commercial equilivalent.
Correct me if I am wrong, but doesn’t the usage statistic on Google’s Zeitgeist register the percentage of traffic which came from (and went to) each operating system?
This is entirely different to the percentage of google user using a particular operating system. FFS, look at what the popular searches were, most of these searches came from people who aimlessly google all day long (where as us Linux users are too busy on support forums, having flame wars and installing distros to waste all our time googling for crap).
All that this statistic shows (IMHO) is that Linux users are: A. More efficient googlers and B. know how to get where their going and don’t need to google so much anyway ([randomPoint]because Linux stuff on the net is better organized[/randomPoint]).
I think the above comment is a very keen observation. I knew there was some way to explain the skew in statistics. Just goes to show you guys, statistics can lie.
I use Sherlock and Google isn’t a resource used by Sherlock. I can’t remember the last time I used Google. I don’t even waste my time with the cute little Google box in Safari..
Kevin Arvin Posted “It would be interesting to see those stats broken down by time of day. I wonder how many are googling from work and therefore have no choice of platform.”
According to the Linux press, the businesses they work for DO have a choice of platform. We’re forever seeing stories on the Linux news sites of yet another business considering or shifting it’s desktops to Linux.
nonamenobody Posted: “This is entirely different to the percentage of google user using a particular operating system.”
Quite right there. People seem keen to make all sorts of analysis of these figures, which is a bit silly really. These figures don’t tell us all that much.
“look at what the popular searches were, most of these searches came from people who aimlessly google all day long (where as us Linux users are too busy on support forums, having flame wars and installing distros to waste all our time googling for crap).”
Which implies that you think typical Windows dummies aren’t switching to Linux, I guess. No growth for Linux?
“All that this statistic shows (IMHO) is that Linux users are: A. More efficient googlers and B. know how to get where their going and don’t need to google so much anyway ([randomPoint]because Linux stuff on the net is better organized[/randomPoint]).”
I can’t say that I see anything to support your opinion based on the Google statistics. It reads like speculation to me.
Here’s another possibility:
Perhaps Desktop Linux is displaying very little growth. People are not switching from Windows to Linux in droves, as some in the Linux press desperately want everyone to believe. Perhaps Grandma is still using Windows to perform her “googling for crap”.
All we can see from these statistics is that Linux OSes make only 1% of Google’s search queries, whereas Windows and Mac OSes make up 95%. I don’t think there is much point in trying to analyze the figures any further than that, because it’s all going to be biased speculation as to what it all means.
All we can see from these statistics is that Linux OSes make only 1% of Google’s search queries, whereas Windows and Mac OSes make up 95%. I don’t think there is much point in trying to analyze the figures any further than that, because it’s all going to be biased speculation as to what it all means.
There is some logic to what you say. But, by your own logic, the numbers provided by Google are themselves speculative, and therefore meaningless.
[insert epistemology argument here]
“look at what the popular searches were, most of these searches came from people who aimlessly google all day long (where as us Linux users are too busy on support forums, having flame wars and installing distros to waste all our time googling for crap).”
These are some hilarious generalizations that have nade ny day, er night…
Oh and btw, I am using Moz/Gentoo-Linux to access osnews, everytime. I have tossed windows xp in the trash, I only have Gentoo on my laptop and my desktop runs in linux all the time but has windows xp on it. I even play all of my games in linux the ones that aren’t native run on wine/winex.
Too bad there is no real statistics about computer usage, only some that are just “guesses”
Once again you see the Zeals commenting. Linux hardly have a share… 1%. Is this when we’ll see any rant like “It’s almost ready for desktop” again?
Linux is not considered to be an option by many and will most likely not to be ever. The only ready thing about Linux is the neckbreaker that will finish off the penguin!
However, it is remarkable what a large share there is of unknown OS’es in usage today… 4%… is that BeOS? can it actually Be that BeOS is larger than unix just that it doesn’t have that many Zeals?
Windows ME must be in that unknown figure somewhere too, and I guess there’s people out there using BSDs on their desktops too. And there’s other older Windows that may still be in limited use too.
Is it possible that some Linux distros (or indeed other OSs) couldn’t be detected properly?
is that most Linux users do not have to go to the Google front page to do a search since both Mozilla and Konqueror and Galeon have integrated searching. If Google is counting based on hits from OSes, this will skew things. But if they’re counting based on user agents, it shouldn’t affect things.
is that most Linux users do not have to go to the Google front page to do a search since both Mozilla and Konqueror and Galeon have integrated searching.
Of course, there are people on Windows using Mozilla (in one form or another), too. Personally, I use Windows (2k/XP) on the computers that I do my web browsing on (anything I put Linux or BSD on doesn’t get a web browser because it doesn’t need one, those computers usually aren’t even connected to the outside world, or would have to be allowed access by the firewall). I use Phoenix primarily, IE for sites that seem to be faster on IE (or have more features I want on IE, interestingly hotmail seems to be faster on phoenix). I also tend to open sites in VS.Net from time to time when I’m working and want to check a site or look for something related to what I’m working on (and, also interesting, VS.Net gives me tabbed browsing with IE6 as the renderer, now why can’t they do it in the stand-alone IE6 browser?).