Windows internet share drops below 90 per cent

The Linux sub 1% has been posted in various 'browser/OS' statistics of late. It doesn't sound implausible but as always the limited view of the article and/or data set begs more questions than it answers.

During the heyday of th IE vs. FF 'war' I recall seeing a few more specific browsing statistics. What really stuck out at the time was the discrepancy between daytime and evening browser usage. During the day IE was clearly the king, with FF trailing behing at 10-15%. After 6 p.m. things changed dramatically, with FF doubling it's share to approximately 30%. Has anyone come across this phenomenon in terms of Linux usage in a representative data set?

Well, I use Firefox exclusively for web browsing, both at work (from a Windows Machine) and at home from a Linux machine, so the dichotomy is clearly there where I'm concerned. On the other hand, i suspect I never, or very rarely hit the sites that are used in these statistics.

Statistics vary too wildly or are too biased to be of any use on most websites. That includes The Register and probably even Google. My stats for November:

Windows: 49.2% Linux: 42.6% Mac: 5.6%

Firefox: 61.1% Internet Explorer: 16.5% Mozilla: 10.6%

That's from about 340,000 page views.

Web browsing statistics indeed don't say too much, but playing with them can still be fun. In your stats, do you see an increase in the ratio of Linux vs. Windows users during the evening/weekend?

@ABCC: I can't say right now. I'm using awstats to generate monthly stats for me. I'll have to check if there is a way that I can query its database with arbitrary parameters so I can split between work days and weekends.

I do see significantly more traffic on weekdays than I do on the weekends.

@ABCC: awstats apparently doesn't track it that detailed. I also have Google Analytics but that is quite limited. I have been able to gleam some information from it though.

Browsers: A very curious thing is going on here. I have roughly the same amount of IE, Safari and other browser users every day. But firefox jumps to a high on weekdays (3 times as high as the rest) and then drops back in the weekends (to about 1.5 times the rest). This is in absolute numbers, not relative to each other.

For OSes it's less erratic but in absolute numbers, Windows beats Linux on weekdays and Linux beats Windows in weekends. Linux is about the same every day while Windows goes up and down.

If you combine that with my above remark that over the whole I get more on weekdays than in the weekend I can draw only one conclusion: On weekdays I get a big influx of people using Firefox on Windows. Probably sysadmins looking at my technical articles (my top ranking article of all time is about configuring Apache and Subversion to use Active Directory for authentication).

Thanks for taking the effort to look it up. It does sound as if what you say is credible, especially on the more technical sites it would seem more likely that people browse more during office hours. Good to hear that Linux use goes up at the weekends too. Even some of the diehard windows admins I know are beginning to kick Linux' tires in their spare time (weekends....)

Sander has become our resident "Answer Man" around here..knows all kinds of stuff and things..

/bows to Sander's extremely wrinkled cerebrum.

Somebody's gotta cause its not me..

;-)

>wrinkled cerebrum

@Scott - an apt description...

Thanks Scott :-) I aspire to be somewhat of a polymath so I try to know as much as I can about a wide array of subjects.

Polymath..can I be a polyword? No, because its not a word!!

Stupid Words..

Polymath, always makes me think of being a Mentat.

Polymath, also known as renaissance man or homo universalis: http://en.wikipedia.org/wiki/Polymath

Do you think they deliberately tried to make the graph look like Pac-Man? Yellow for IE, dark gray for Mac on a black background and the rest of the pie wedges at the bottom.

Knowing El Reg, I'm quite positive that it's intentional :-)

Nonetheless, you must admit it is much, much better than this: http://marketshare.hitslink.com/os-market-share.aspx?qprid=9

If the numbers were not there, one would have to wonder why bother? Moreover, when splitting fractional percentages, what is the estimated error? Remember this is a survey with many variables, not a precise measurement of hard data. My guess is the error range exceeds the average observed monthly change.

Personally, I cannot wait until the dentist enters and orders open wide, so we can finally see the teeth.