Recently we've had a couple of articles and discussions on OSNews about browser statistics. Every now and then someone in the comments will ask what OSNews statitstics are like, and what browsers and operating systems our readers are running. Well, recently we installed a fancy new web site analytics program, Mint. I figured I might as well put it to good use by showing some of the browser and platform statistics we've gathered over the past month.
Post a Comment
Member since:
2006-08-17
Member since:2008-04-18
Member since:2005-06-29
Member since:2008-02-26
Member since:2005-07-07
Member since:2005-07-06
Member since:2005-07-06
Member since:
2009-01-20
Member since:
2006-12-22
Great statistics from this site.
My girlfriend uses Firefox after a long lobby... so I think the world is starting to change. I will ask her to read OS News to make clear the statistics!
At least seems balanced for technical persons that do testing stuff or work one. About IE, I think is the case that someone told here: at work most we cannot upgrade our browser, as we are in a corporate institution. At least to be sure that are patches around...
Member since:
2005-07-24
Member since:2005-07-15
I'm also an Epiphany users. I have just seen that Epiphany identifies itself as "Mozilla/5.0 (X11; U; Linux x86_64; en; rv:1.9.0.6) Gecko/20080528 Epiphany/2.22" so i think most of the mozilla-users are in reality "gecko-based-browser"-users.
BTW: I'm surprised tha Chrome has already almost catched up with Opera.
Member since:2005-06-29
.
Member since:
2007-07-10
If you could compile a list of the "others", I'm sure people here could fill in the blanks so you could update your script/software you use to generate your statistics.
AWeb, IBrowse and OWB are Amiga browsers.. OWB, which stands for Origyn Web Browser has become one of the most popular one in a short time.
I'm sure there's tons of others.. if you supply the unknown, perhaps we can make it known?
Member since:
2006-12-27
Member since:
2008-12-05
From the article:
"As it turns out, most of our readers have Flash installed, with only 8% not having Flash ready on their systems. Unsurprisingly, Flash 9 and 10 are the most popular, with respectively 24% and 67%."
Your stats don't mention Swfdec or Gnash at all, which I would expect to be in use by at least 1% of your visitors. I'm guessing that Swfdec and Gnash users are lumped into the 8% that don't have Flash (which would make sense if you had specifically said "Adobe Flash Player", but is a bit ambiguous without). Is that correct? Do you have any stats for Swfdec and Gnash?
Member since:2005-06-29
Member since:2008-07-15
Member since:2005-08-29
Member since:
2009-03-20
Member since:2005-07-07
Member since:
2008-12-05
Member since:
2007-04-04
Member since:2005-06-29
And...?
Whether you use it at work or at home, you're still using it. whether you use it by choice or not, you're still using it. Every hit is as valid as another
.
Member since:2005-11-10
Member since:2005-07-26
I think a good point to make, would be that many people don't use their preferred OS or browser at work. I think the stats would be way different, if you could exclude that part.
For instance: Exclude hits from the US in working hours and the hits from Europe in working hours. Can you do that?
Member since:2005-06-29
Yes, and car sales statistics would probably also be a lot different everyone could just pick the car they wanted, instead of the car they can afford.
Wat I mean to say is is that it makes no sense to remove the work hits form the stats. Hits are hits are hits. These statistics have nothing to do with what people want to use; they show what people are using.
Edited 2009-03-20 21:46 UTC
Member since:2005-07-26
Member since:2005-07-07
Ehhhh.
You're very wrong Thom. Filtering out stats is telling you something about the users of this site. So cutting out daytime to check how many of those hits are work-related actually IS interesting.
No matter how many stupid far fetched analogies you come up with.
What a cheap and very low-brow reply from you, I wouldn't expect that!
Member since:2009-03-20
When I am at work, I am compelled to use Microsoft and Internet Exploder. At home, I am free to use Linux. I will check the osnews site from work on break using Microsoft, but in my heart I am using Linux. So, even though when I get home and I check in using Linux, I still have the stain of having used Microsoft for probably half of my visits to your site.
Would it be possible to factor *my* Microsoft/IE hits to be Linux/Firefox ones? Let's say I average 1 per day - that's 200 per year. After all, if I had a choice, I would have used Linux over Microsoft.
Would that make a difference in the stats? If more of the so-called Microsoft users were like me - using it because we are obliged to at work - would that make a difference?
After all, you are presenting your results to advertisers. I don't want them getting the wrong idea about my computing choices and interests and then making marketing decisions based on that. Just because I use Microsoft half of the time I surf to osnews does not mean I am interested in anything that would pertain to voluntary Microsoft users. I would still be much, much more interested in and likely respond to OSS/Linux related items than in Microsoft ones. However, the stats to not reflect this. Both Microsoft and FLOSS providers may incorrectly choose to either target or avoid me as a result.
To prove the point, this is typed on a laptop that I bought from a Linux-only vendor, in Linux of course. I specifically looked for a provider who offered Linux as a choice, and discounted ones who did not offer Linux or made it difficult to select Linux. Had those latter advertisers thought that my Microsoft-based hits on osnews indicated that I was a potential customer, they were misled. I even researched my choice at lunch times from work, using Microsoft. Taken at face value, that would have seemed to be a contradictory scenario. In the context of my work usage, however, it does make sense.
Member since:2005-07-06
You can do that, change your useragent to Firefox on Linux
(interesting twist huh?)
Member since:2005-07-06
I'm not allowed to hook up my Macs at work but found a work around. I work in Tech Support as a Systems Analyst and we have DSL line in our department so we can test VPN accounts which has a wireless router connected to it. The tech that manages this router was sympathetic and gave me the key so I can wireless connect to it. So my stats are based on my G4 PowerBook which runs Tiger.
I notice that it didn't say anything about Safari 4 beta. I use that but before that I used the nightly builds for Webkit. I don't see that listed either. I guess it counts as Safari.
One thought, how about showing us the stats for domains for our e-mail addresses that we use to register here at OS News. That might give us more of an idea where our loyalties lie. Yes I realize that some people will have comcast or some other cable/dsl company as their e-mail address. Those like me those have an @mac.com (@me.com - I HATE me.com, it is so stupid. "me me me me me". I really not that into "me".
Member since:
2005-07-15
Member since:
2006-02-05
Member since:
2006-05-09
Member since:
2006-12-05
Why not make this crime punishable by not displaying the page unless another (any other) browser/IE version is reported?
A simple browser upgrade is long overdue, and seriously, people who still use IE6 *should* be punished... it's been eight years since that atrocity has been released.
Also, Firefox has no exclusive control over the browser market, nor do they lack competitiors, so "big bad evil monopoly" sounds overkill. They're not even a monopoly, as their 54% and the presence of several other browsers on the pie chart show.
My stats tend to be (currently):
-Linux (Ubuntu 8.10)
-Firefox 3.0.7
-Flash installed, but almost always blocked through the use of Adblock Plus and NoScript
-Screen resolution of 1680x1050 (browser almost never maximized)
I have pretty much all the major browsers (and some minor) installed, but for a while now I've been using Firefox exclusively. I actually have Chromium, Dillo, Epiphany*, Firefox*, Konqueror, Midori, Netsurf, Opera*, SeaMonkey*, lynx, w3m, and links (ones market with * are the ones I use most). By the way, the SeaMonkey 2.0 alphas are looking really nice...
Member since:
2009-03-20
Member since:
2005-07-06
Stats can always be manipulated to say whatever you want, we all know this.
This example is pretty useless, I browse the site on my main machine, and my secondary laptop, and my work machine.
My main machine dual boots between Ubuntu and PC-BSD, my laptop between Ubuntu and WinXP and my work machine between XP and Vista.
So I will be showing up with a few skewed results.
What would be better, is, if Thom can come up with a poll that asks what our preferences for a Primary OS and Primary browser are.
After all, in work, I WISH I was using something else.
Member since:2005-06-29
After all, in work, I WISH I was using something else.
Uhm, this isn't about what you WANT to use. These statistics are not meant to look into the minds of people and find out what they WANT to use.
The statistics show with what platforms and browsers people access OSNews. That's all they do. Nothing more, nothing less. They are not useless. You find them useless because you expect these statistics to give you information on something they cannot give you information about.
Similarly, a poll, as you suggest, does not provide us with the data that the statistics provide us with. They measure two completely different things.
Is the difference really that hard to understand? I'm intrigued.
Member since:2005-07-06
Member since:2005-07-07
Member since:
2006-12-26
Member since:
2005-12-18
Thanks for listening, Thom!
Now, three things.
1 is for you, Thom: What timespan do those statistics represent? And did you specify an address we could go to see them when we wanted to (updated, of course)? Don't remember if you did. I'll reread the article later.
2 I fall into "others" as I read OSNews mostly from akregator (konqueror based) though I use firefox as my browser of choice in my box (like anybody cares, right?).
3 Well, I think this is a little more "empiric" data confirming my thesis that IE has really lost the hearts of IT people. http://maratux.blogspot.com/2009/02/has-ie-lost-hearts-of-it-people... I guess we'll have to wait and see how IE8 does during the following weeks/months.
Member since:2006-12-26
Member since:1997-10-01
Member since:
2009-03-20
The "interesting" statistic about the number of people still using Firefox 2.x can probably be explained by BeOS/Haiku users. Firefox 2 is the last available version for BeOS/Haiku. I'm using Firefox 2.0.0.18 on BeOS.
(I've also visited here using IBrowse on an Amiga 500.)
Member since:2006-10-28
Member since:2008-07-07
Member since:
2005-07-06
Member since:2005-07-07
Member since:
2008-12-16
Member since:
2009-03-20
Anyone running a non-standards compliant browser (I.E., IE) should be ashamed of themselves.
I've been running Opera steadily since '96 and none of the others measure up. Overall, it is the most non-geek user friendly browser out there.
BTW, All you Open Source, Free software geeks, keep on rockin'. It amazes me that people whom have a passion for coding and freedom; pose the biggest long term threat to Monoplosoft
Sincerely,
non geek.
Member since:2006-10-08
No, because those (non-geeks, "average users") who usually use IE don't notice it. They're not using a particular web browser, they're "using the Internet". For them, it is "Windows" and there's nothing else.
I whish I could agree with that. When we would talk about Opera versions of the past, v5 or v6, I would surely say yes, it's fanfastic. But today's Opera, while still being a very powerful, standard compliant browser, rich at features, and a joy to use, has a growing taste of slowliness, complexity overhead regarding configuration and... I know it's a bit complicated to say, but Opera got slower and more complicated to use. While some changes really appeal to me, others get on my nerves, so I'm observing myself more and more often to be running Firefox - the browser that's always been accused of being "too slow", but in fact, it's sometimes much faster than Opera and displays pages that Opera don't seem to be able to display. But when I use other browsers, I always know why I miss my Opera. At this point I would not want to change, allthough I'm always be told that Firefox can be set to use mouse gestures as well...
Don't get me wrong: If a page is standard compliant, Opera has no problems, so I'm happily using it for OSNews. I even don't change the user agent identification string.
(Erm, Thom, do the statistics take into account that a Linux / UNIX browser can change it's identification string to an "Internet Explorer" pattern?)
Another nice effect of using Opera with OSNews and the pages it links to: When "Flash" is used, you just see an empty box. No dialog window annoying you to download some strange plugin, very relaxing... :-)
Ah, by the way, non-geek FreeBSD + Opera (web) user here. =^_^=
Member since:2007-04-05
I would like to note, that after many years being advanced Opera user You simply can not migrate to anything else. As a web developer I have all the browsers installed (some of them in virtual machines to avoid conflicts) and I use them frequently. But experience is always awful, because their interface is like a brick: you can grab it and throw somewhere, no rubber handling, coffee-making etc.
So if you are addicted to Opera UI and its features (proper mouse gestures, proper keyboard-only interface, proper tab handling, proper shortcuts and many other things properly made) there is no escape. I do not want use a brick instead of a browser even if brick will kill my enemies with ease.
Member since:2005-09-27
Member since:2005-07-06
I think we can safely assume that no one visiting this site with IE 6 is doing so out of their own free will. Many of us have jobs and many companies still have IE 6 as their default browser. The company I work for only approved IE 7 as optional upgrade for those who want it a few of weeks ago. Any browser other than IE 6 or IE 7 is not allowed unless you are working on a web related project and you need other browsers for testing.
Member since:
2007-08-24
Member since:
2008-10-09
Member since:
2006-01-09
Member since:
2007-03-14
The 90% of people visiting this site with Flash installed should burn in the hottest fires of hell.
I'm a part of that "Other" 2% (*BSD surely has a larger percentage..)
I'm a part of that 56% (Firefox 3.. sigh)
I'm a part of that 8% (No more f--king flash!)
I use 1024x768.. f--k people with high resolution LCD screens.
Edited 2009-03-21 05:34 UTC
Member since:
2005-07-06
It would have been interesting to see any differences in statistics for working hours and non working hours. This is of course a bit more difficult to do, because you need to take different time zones into account, but non the less interesting.
My guess is that the high figures for ie6 reflects people browsing from work where they often have no choise in what browser to use.
Member since:
2005-06-29
Member since:
2005-07-06
Well, I wouldn't say it's surprising either. E.g. I'm reading the site from Linux at home, but at work I mostly have to use Windows, thus I read OSNews more from Windows than from Linux since I - unfortunately - spend more time at work than at home on workdays. And in Europe Macs are much lower in number than in the US, just an example, we have around 300 people where I work, and there are no more than 2-3 Macs, desktops and netbooks included.
Member since:
2005-07-12
I know I've been a bit harsh on the site since the redesign in terms of things like fixed metric fonts, fixed width layout and accessability failings, but there is one important thing that makes these numbers more relevant than a LOT of statistics presented out there...
Unlike most stats, these are based on a website that actually WORKS in all browsers. As a web developer I recently got into a spat with another developer over a site we were working together on, where he was saying we didn't need to worry about Opera and Safari compatability because the existing site didn't see more than 1% total traffic from either... When (duh!) the existing site didn't WORK RIGHT in either. If the site doesn't work right in part of your demographic, guess what? You're statistics are going to be skewed. It's like doing approval rating stats for Obama at a KKK meeting, or Bush approval ratings from members of PETA - your numbers are going to be a little off 
A great example of this is likely the browser statistics for gmail or Microsoft's websites. "under the hood" Opera defaults to "mask as firefox" on those sites, so their developers never get to even see how many Opera users they have, lumping them together with FF users in their stats. (neither site working properly in Opera without the user javascript effectively rewriting the page!)
I'm actually suprised to see your most common resolution so high - the most common of the sites I maintain still ends up 1024x768 right now... though I'm probably one of the few listing as 1920x1200... Given the fonts being fixed size and the average user being lazy, the fixed metric fonts likely limit the resolutions you get of common visitors, most blowing right past the site instead of doing I do, and zooming everything in 40% and hitting 'fit to width' every time I visit the site.
One thing that REALLY suprises me is that the IE6 slice of the pie is so low - with so many people forced to stay on it by their workplace, and the majority of users visiting this site not being DUMB ENOUGH to use IE by choice, IE6 should be more prevalent than 7 given the number of IT departments that STILL won't touch newer versions with a 20 foot cattle prod or cannot upgrade to it because of some rubbish decade old in-house crapplet built around Trident and ActiveX.
The IE 5.x demographic probably is NOT coming from win95/98 as much as you'd think, since IE6 will run on those platforms quite well... I'd be interested in knowing how many of those IE users are actually using IE 5.2 on the Mac, since a LOT of mac users still use that outdated version since many banks and financial websites ONLY work in IE, and that's the most recent available to them. I also know people using MacOS 9 on G3's or less in which case IE 5.2 is the best browser available to them since iCab is a joke, and the last gecko based browser was mozilla suite 1.2.
Older copies of Windows CE or Mobile also likely count towards the IE 5.x demographic since it's often impractical to upgrade those types of devices to run newer versions of the browser.
But that's the fun of statistics, the numbers are concrete, but you can make them say almost anything via differing interpretations or ommissions.
Still, DAMNED good to see a website where IE (curse it to hell) is the minority even if the large slice of firefox likely being more from lemming factor than intelligence.
Edited 2009-03-21 09:24 UTC
Member since:2005-11-15
Member since:2005-06-29
Well, why didn't you send us an email with screenshots and information? We can't fix something if we don't know about it! We are not rich enough to buy Amigas or MorphOS machines, so we have to rely on information from others.
Member since:2005-11-15
Ibrowse does not support CSS, would you change your web site so it doesn't use CSS?
If the answer to that is yes, that I'll send you a screen grab. But that's not the answer I'm expecting to be honest.
Anyway, I though that at least one of you had a Peg II? Morphos is free (as in beer, for 30 minutes anyway), iBrowse is also available for 68k classic Amigas.
BTW I'm writting this in Opera on PPC Debian Lenny
Edited 2009-03-21 14:27 UTC
Member since:2005-06-29
Member since:2005-07-12
While it's not a be-all end-all solution, this is where semantic markup comes into play - really non-css browsers are being told to **** off at this point, with the majority of developers moving to CSS only layouts since one of the big approaches (and one I heartily endorse) is called separation of presentation from content. Anything in the HMTL should say WHAT elements are, not how they appear - appearance is what CSS is for.
BUT, hand in hand with this separation is another technique called semantic markup - something that (as I mentioned in the past) is missing from this incarnation of OS News. One of the advantages to semantic markup is that when you don't have CSS available you usually at least get meaningful formatting on the page. You have a heading to a section, use a HEADING tag NOT a DIV. You have a paragraph, put it in a PARAGRAPH tag, etc, etc. Quite often using meaningful tags in this role also means a hell of a lot less classes and ID's, meaning less markup... and remember one of the goals of modern markup techniques is to get as much as possible OUT of the markup to save bandwidth.
Take the main page at OSNews, where right now (on top of the buggy confusing formatting - and don't give me that "it's in a CMS" excuse) You have an outerdiv with the so descripting class "item2", followed by a .newstitle1 and .newssubheader1 div, that star thing that I've never been able to figure out what it actually does, but should probably JUST need the anchor to function and not an anchor + span + image with a image replacement technique, an obvious paragraph, etc, etc.
Since we can't post code in here, I'll just slap together a quick example on my server.
http://www.cutcodedown.com/for_others/osnews/template.html
Which shows exactly why slapping div's willy-nilly around everything is often no better, and in many cases worse than using a table - when BOTH are in fact the wrong approach.
Member since:
2007-01-24
We always see here and in other websites news about US-based statistics of OS market share and in almost all linux percentage is much lower than MacOS X. But you can observe that OS News has more linux readers than MacOS X readers, even using the english language and therefore be accessed more by americans.
But I always suspected this is not true for the entire world ans specially outside USA, where Macs are popular.
I think that linux has now more market share than MacOS X globally and this difference is very big in developing and european countries.
I'm a proud desktop linux user and I never understood how an operating system which runs only on a proprietary hardware can be more popular than a free (as beer and as freedom) operating system running on the same PC hardware used by windows. Even considering that linux is wrongly known as a geek/nerd OS, Apple has not suficient presence nor competitive prices to has more market shar than linux.
Edited 2009-03-21 15:07 UTC
Member since:
2005-07-06
Member since:
2006-03-02
Member since:2005-07-12
1) I believe they fully recognize that the numbers are skewed. OSNews users tend to be a bit sharper than Joe Sixpack.
2) A "spiritualist" website is going to have numbers just as skewed - just the opposite direction since the majority of technical minded people tend to not put much stock in that sort of thing. (trying to put that as politely as I can and not go all MC Hawking on you)
3) You know that with 201 validation errors on your page you do not even have HTML, you have complete gibberish, right?
4) Tables around single TD? Empty heading tags? Nonsensical heading orders? Hitslink markup inside your HEAD instead of the body? Inlined CSS? Presentational Markup? Marquee??? 30k of HTML for 4K of content? No wonder the layout is broken in Opera (though that could be large fonts compatability) and you are probably driving away business with the 832 megabyte website in some 33 files - at least 30 seconds on firstload here and I'm on a 22mbps downstream.
5) Three separate stats? here's a tip, pick one - or even better, use the ones built into yer damned server and save end users the hassle of loading three separate javascripts from three separate servers increasing the page load time exponentially. Reminds me of a buddy of mine who had a slow as molassas page, so he kept ADDING more statistical tracking to it trying to figure out why it was so slow when his own server had little or no load...
6) Let's just say, I wouldn't go bragging about that site.
Edited 2009-03-22 05:59 UTC
Member since:2005-11-15
It's also based in the USA, If I as a non-American running on a non-Windows OS, wanted a salt lamp, I'd take one look see the dollars and go off and find a UK based one.
So you're going to have a Northern American bias there. Which could, and I only say could explain the high Mac/low Linux ratio.
But I think your results are a lot closer to the average than the stats for this site.
Member since:
2007-04-05
Member since:
2005-07-06
Member since:
2005-12-09
Member since:2005-07-06
Member since:2005-12-09
Member since:
2009-03-23
It would be very interesting to look at CPU architecture statistics. user-agent strings contain this data (for example, mine is "Mozilla/5.0 (X11; U; Linux ppc; en-US; rv:1.9.0.7) Gecko/2009032022 Iceweasel/3.0.6 (Debian-3.0.6-1)"), but I haven't found such kind of stats anywhere, there are only OSes and browsers.