It’s been a while since we’ve done one of these, so here we go: some detailed platform and browser statistics for OSAlert. They’re collected using Google Analytics, between 12 April and 12 May. As always, these statistics are only relevant for OSAlert, and can, in no way, be extrapolated to any other site.
Windows is still king, with Linux as a solid second, followed by OS X in third. Android is only slightly ahead of iOS, but even when combined, the post-PC world cannot beat the third most popular PC platform. Also of note: Windows Phone barely even registers – in fact, FreeBSD is almost three times as popular as Windows Phone, and even Solaris is just as popular as WP – and we have one person browsing OSAlert on a Series 40 device (rock on!). Chrome OS isn’t really very popular among OSAlert readers either.
Moving on to the browser statistics, Firefox and Chrome keep each other in balance, with Safari a distant third. Internet Explorer is close to irrelevant, as is the Android Browser (the AOSP one; Chrome for Android is added to Chrome). Interestingly enough, the Origyn Web Browser registers (barely, but still) – which is popular mostly among AmigaOS, AROS, and MorphOS users. As a final note, Sailfish OS might possibly register as Maemo; I’m not sure.
It’s always interesting to look at these statistics, as it gives us an insight into what kind of stuff you men and women are using. I’m happy to see that you are still a varied bunch.
Can anybody who uses Safari explain why it is superior to Firefox/Chrome?
I was using Chrome until the official Java releases for Mac became 64-bit. I now use Safari.
There were some work-arounds to get Chrome working with out-dated Java builds but the work-arounds didn’t seem to persist over time. Also, at some point, it just seemed wrong to continue to have to use an out-dated 32 bit version of Java just because Google wouldn’t get off their ass to release a proper 64 bit Chrome for Macs.
P.S. There’s 1 site I need to access for work that requires a proprietary Java plugin in order to log into the system. Otherwise, 32-bit Chrome with no Java integration would be fine for me.
Edited 2014-05-13 21:06 UTC
I switched to Safari when version 7 came out. Performance is top-notch.
I’ve had some problems with Firefox (lags, resource-hog), and since all the browser add-ons I need are available for Safari, I just quit using Firefox altogether (well, almost, I’m a web developer
.
Chrome is fine, but certain OSX specific functions (gestures, etc.) don’t work as well in chrome as they do in Safari (e.g. pinch-to-zoom). I use chrome for work because I love the developer tools and Netbeans integration, but I prefer Safari when casually surfing.
I use Safari because:
– I prefer it’s text rendering to other browsers (more accurate)
– It integrates well with the rest of OS X (and iCloud)
– I like the gestures (not all are in other browsers)
– I like the highly polished and Mac-like UI
In addition to the two comments above, I’d add that Chrome on Mac OS X seems to have this odd quirk where Cmd+R and/or the refresh button will stop working. That one specific problem combined with other issues (mentioned above) with Chrome makes Safari more reliable, better performing, and more Apple-like.
Chrome was bloody slow on my Mac, same for Firefox. The latter has since improved and I’ve removed the former from my system. Two other advantages of Safari over competing browsers is Reading List (tap a button and an article is automatically saved for later) and the syncing of open tabs in Safari on OS X and iOS (which I use mainly for opening recipes on my iPad in the kitchen from an open tab on my Mac).
While on this subject, I have to rant…
So I generally use Firefox everywhere (call me old fashioned, I used to love Netscape when it was the king). But unfortunately there is a site that I use on the intranet at work that doesn’t work correctly with Firefox. So I had to look elsewhere, and we are given Macbooks at work.
So at first I tried Chrome. I really hate that Chrome, no matter where it is, always decides to look like itself and nothing else. I have a nice wood textured theme, and Chrome just looks horrible in it. That and I always feel like everything I’m doing is sent to Google…
So then I decided to give Safari a try. Hadn’t ever really used it before. But now I know why. Whomever keeps saying it’s fast… maybe mine is broken or something, but it’s slow as crap. I feel like I’m using Netsurf on my Amiga whenever I use it.
To describe what Netsurf does… it seems slow because it likes to load everything in memory, then display it. Seems Safar does the exact same thing. So pages all take about 15-30 seconds to load (same as on Netsurf). Drives me nuts. The loading bar on Safar starts, goes about a 10th of a way, then sits there… then boom, all the way to the end.
So yeah, I’d stick with Firefox if I could, I don’t see the draw with Chrome and Safari. People have complained to me before that Firefox is a pig. I had about 200 tabs open one point, and it was still only about 600mb of ram. I think the the problem some have are usually because they have tons of flash crap open, or indeed don’t have adblock on. In this day and age (and I know web page admins will cry) it’s almost required to have adblock.
Remember the days when ads would pop up in separate windows? Everyone hated that, so all the browsers introduced pop up blocking… so now all the ads are in the same damned window as the rest of what you’re trying to read, which to me is worse in a lot of ways, at least with pop up ads, you could tuck them away behind the main window.
I noticed earlier that one of the forums I visit wasn’t fully blocking the ads, and it looked like after every other legitimate post, there was an advertising post…
It’s freaking ridiculous! Duckduckgo is one of the few pages that do it right… have a single ad or two on the side bar that’s a non-intrusive (read not flash) ad and you’ll have people that may even click on them, and they’ll disable adblock.
Maybe if someone created a plugin that would rearrange the ads to make them do that, and include flashblock in with the mix, it’d be a win for everyone…
But if you have adblock, you don’t see which sites respect you… (by moderate use of ads)
On some platforms (OSX) it starts much quicker.
Also on OSX it zooms in and out smoothly with mousepad pinches … though the Chrome may have caught up a bit.
Superior would be a strong phrase – they each have their nuances.
Happy to be one of the few Windows Phone users
(that is, as long as I’m using WP, I’m not unhappy about it – even though my next phone will probably be Android again)
WP8.1 FTW! (and FFS, get a real dropbox app and an ING bank belgium app!)
I reckon that the stats for mobile browsers are higher than reported, because many browsers are redirected to the mobile version of osnews, that probably does not contain the tracking code for this statistics engine.
Good Point.
That used to happen quite often to me. I’m not sure how I got past that, but I haven’t hit the mobile website in a while.
That might be true – but the only browser I’ve seen recently to do that was one of the Opera ones. iOS and stock Android (Nexus, so Chrome) both give the full site.
Windows Phone (8.1) gives me the mobile site as well, even though I used the setting that should give me the desktop version “always”. It is driving me nuts because the comments are only shown 10 per page with no respect for their score (I normally only read level 3 and higher)
Or, the stats for mobile browser could be under-reported as most of the mobile browsers (Chrome, Firefox, Opera) are configured to request the desktop site by default. I believe the AOSP Browser is the only one that defaults to mobile, and can’t be set to desktop on a permanent basis.
I know all my mobile browsers (Chrome and Opera Classic) are configured for desktop.
The stock android browser (AOSP) will by default fetch the desktop version of OSAlert. It will fetch the mobile version of Wikipedia, unless configured to fetch the desktop version. It depends on the site in question whether a mobile or a desktop version is served. Particularly confusing when using a tablet, sending you to mobile sites from time to time, semi-randomly.
Interesting. I haven’t used it in a long time. Last time I used it, I found that it presented the mobile site by default, and you had to bring up the menu and check the box for “Request Desktop Site” to re-load the desktop version of the page. And you couldn’t make that a permanent setting … you had to re-toggle it for every … single … page … loaded. One of the main reasons I moved to Opera (back before it was considered Classic) on my phones (more recently moved to Chrome).
I read via Feedly so I rarely show up directly on the website.
So, I was curious where Haiku and WebPositive might land on that list… since I suspect GA doesn’t recognize either of them.
I looked at the GA stats for haiku-os.org, and I notice that we have a large number of visitors with an OS of “(not set)” – so I suspect that is how Haiku is detected when using WebPositive to access a site.
Furthermore, I believe WebPositive is being detected as Safari based on the stats I saw there.
Just some interesting observations for those who are curious
Yep, that shows up as one of the scripts blocked by ghostery
Seriously, on this site I’d expect a large proportion of users are blocking google’s trackers, no? Nothing is as accurate as tracking the information directly from the server hosting the website.
Well, at least you would have a lot more information to work with – since GA generalizes stuff into large buckets, and alternative OSes don’t tend to fall into them very well
That’s a good point. I block GA in all my browsers. If sites want to track me, they can bloody well do it on their own hardware, using their own software, and not shove extra, useless JavaScript onto my machine.
It’s an interesting list and I suspect not very representative of the wider Web!
I don’t know whether it’s helpful (I’m not familiar with the machinations of Google Analytics) but the user-agent strings generated by Sailfish are the following:
Default browser: Mozilla/5.0 (Maemo; Linux; U; Jolla; Sailfish; Mobile; rv:29.0) Gecko/29.0 Firefox/29.0 SailfishBrowser/10
WebCat: Mozilla/5.0 (Maemo; Linux; Jolla; Sailfish; Mobile) AppleWebKit/534.13 (KHTML, like Gecko) NokiaBrowser/8.5.0 Mobile Safari/534.13
Perhaps these can be used to figure it out?
Would probably have to ask the GA developers
With Haiku, we ran into all sorts of weird UserAgent string issues a while back – where google.com all of a sudden started delivering the mobile version for no apparent reason. We moved a few things around in the UA and it magically started working again. Whatever regex magic they use, it seems to be easily confused.
Science is able to reverse engineer the perceptible universe to figure out what happened 10E-37 seconds into the big bang 13.82 billion years ago, but reverse engineering GA’s UA algorithm is a step too far
Seriously though; you’d have thought there must be some way to reverse engineer this kind of thing, but from what you write, Google haven’t tried to make it easy (I naively assumed there might even be documentation about it!).
Well actually, it’s probably easier than we’re both thinking. I mean for example, I used the haiku-os.org GA stats to extrapolate how I thought GA was handling the UA from Haiku’s WebPositive.
You could easily setup your own website, and attack it from your Maemo device and then review the GA stats that it produced.
I liked your suggestion, so gave it a go. Sadly, although I could see the page impressions register in the real time view, over twelve hours later there’s still nothing appearing in the browser stats
Yet it’s got 4 times as many users as Windows Phone. LOL
Surprising, especially looking back at the few Chrome OS related articles where majority of comments said they hardly ever spotted a Chrome OS laptop in the wild.
Edited 2014-05-14 01:29 UTC
I use Windows Phone regularly, but prefer browsing on Desktop O/S generally. I expect this impacts all mobile browser usage – many people prefer browsing on larger screens.
That “one person” using Symbian S40 would most likely be me. I use Opera 4.5 on a Nokia 3110c (2008 vintage) sometimes, just to mess with Thom’s statistics
I’d say OSAlert site is quite well readable on the tiny display (as of today’s standards). Good job there, guys!
Osnews’s mobile site works perfectly fine with N-Gage QD and opera as well. Great design!
I used to browse the mobile site on a Sony Walkman w580i feature phone using Opera Mini (Java Micro version). Could only see 1 or 2 comments at a time, but was perfectly readable and browsable. OSAlert was one of the few sites that actually worked well with that phone+browser.
That was … 4? years ago? Between 3 and 5 anyway.
I wonder how my N9 registers. I was thinking it could be misdetected as S40, but I guess it is misdetected as something else instead. At least Nokia Browser triggers.
Yep, I browse it occasionally on my N9 too, wonder if it counts under “Nokia Browser” or what… I guess the OS might be too insignificant to have shown up there at all, or maybe it’s lumped in with Linux? Must do some investigation when I go home…
I can we wrong, but if I remember good s40 isn`t Symbian. Symbian is s60.
You are correct, S40 is “Nokia OS”
I’m one of the IRIX guys, although it’s been a while …
I also visit the site in links regularly. It renders really well and is one of the most usable sites out there in links.
Edited 2014-05-14 07:31 UTC
The problem with using something like google analytics is that its javascript based and relies upon third party cookies.. People with javascript disabled, non javascript capable browsers, third party cookies blocked, privacy plugins etc, and the proportion of people doing these things is likely to be somewhat higher on a technically oriented site.
Right, that’s it. I’m going to have to dig out the RiscPC and get Netsurf to ping OSAlert every couple of minutes for six months before you do the next round of stats.
____
EDIT: Although as bert64 points out above, non-Javascript browsers wouldn’t be picked up. Netsurf is (currently) without JS support, so I wouldn’t be surprised if Risc OS had a presence after all. I recommend using awstats on the server side next time around.
Edited 2014-05-14 10:51 UTC
I suspect I’m a large chunk of that FreeBSD share – it’s what I run on the workstation at, well, work. Several of the things I do involve waiting, so …
I’m in the FreeBSD camp as well. Home workstation is running PC-BSD, and I do the occasional browsing on there. It’s generally just a file server, these days, though, so my contribution to the FreeBSD browsing stats is declining over time.
Along with someone else’s comments regarding having Javascript and cookies blocked messing with the statistics, there are also those who spoof their browsers for whatever reason.
I use OWB on AmigaOS, but I have it spoofing as Safari because there are some websites that refuse to work properly when it declares as OWB. Gmail is the one that springs to mind, and there are others… So, I’m probably in your statistics there but unfortunately not in the correct category.
Edited 2014-05-14 12:36 UTC
I typically visit OSAlert using QupZilla and was interested to see it not represented on the list. I wonder if that is because so few people use it, or because it gets identified as something else? QupZilla sends an odd ID string which identifies it as Mozilla, Safari, AppleWebKit and QupZilla so an automated counter might add QupZilla to the Safari category.
I suspect a lot of the IE traffic is from people at work who have no choice of broswers. My personal machines all block Google analytics. I wish I knew the real, un-skewed numbers. Not much you can probably do about it though.
Yep, I’m using IE right now because I’m in work and have no choice. At home it would be Chrome or OWB (spoofing as Safari)…
That’s a really sick requirement. Is that a government work that devises such limitations?
Haha, unfortunately it’s normal company behaviour… And I’ve no mobile reception in the building so I can’t use that either
I thought Opera would have better numbers in these “geeky” statistics.
Actually I wonder if Opera 15+ (Chromium based) is counted as Opera or as Chrome.
Edited 2014-05-14 16:51 UTC
Don’t forget that Opera dropped support for Linux and FreeBSD when they switched to their new engine, so anybody running on those platforms probably isn’t using Opera anymore. At least not if they care about running patched software.
Actually, there was a release of Opera Classic a few weeks ago, I imagine to patch things up a bit…
I wonder if more people here were using it back when it was still a flexible, customisable and feature rich browser? I certainly did back then…
Dumbed down Chopera certainly doesn’t seem to have set the browser world alight…
Guys, congrats and thanks for providing us these stats; anyway, do you think you could provide us with more historial info to see trends or something like that?
^A!Saludos!
IE is my default browser.
I know there are some people who come here who have Amigas. We should get a bunch of us (with WinUAE or real) Amiga freaks and browse OSAlert a ton so we can get on the list.
I surf with MorphOS on OSAlert.com
Using OWB…
the average OSAlert reader is not Joe average internet user.
In particular the OS numbers differ significally.
For me that simply translates to: ask the customer!
Greetings,
pica
Is yours truly
Browser: Nokia5130c-2/2.0 (07.95) Profile/MIDP-2.1 Configuration/CLDC-1.1 UCWEB/2.0 (Java; U; MIDP-2.0; en-US; nokia5130c-2) U2/1.0.0 UCBrowser/8.9.0.251 U2/1.0.0 Mobile UNTRUSTED/1.0
I’m one of the one in every 10.000 who uses OWB (in MorphOS). I was hoping to see MorphOS there in the statistics since I do check this site every now and then and I don’t spoof. But it’s nice to see that we OWB users are, in a sense, over represented among the users here. This is in no way a surprise since people who uses a weird OS like MorphOS are going to be interested in OS development in general.
Edited 2014-05-15 21:44 UTC
When I`ve read article, I thought – haha, I`m the s40, they talk about me!
Because I`m very often reading osnews with my 7 years old Nokia 5310 Xpress Music! But then… oh…. Not only I think, that I am the One. So now – what`s up? Does your stat-software count only Nokia Browser as s40? I use Opera Mini (the Java mobile!), so maybe it isn`t counted as s40? How does it look?
I’m worried about OS News. We see few original articles here any more. Increasingly, it’s just an aggregator site. Meanwhile, Alexa statistics on website visits are dropping (no surprise, given that the content’s getting thin).
What’s happening?
1
Browser: Mozilla/4.0 (compatible; Synapse)