That's not a gain in marketshare, that's some random statistical noise.
Take a look at the graph[1]. IE6's segment is generally trending downwards, but the "wiggle" from month to month is a bit bigger than the trend. This month it happens to be wiggling up. Next month it will probably wiggle down, and ars will get to write a breathless article on the impending demise of IE6.
The correct take-away is that IE6 is declining, but slowly. So slowly, that on a monthly timescale, the change is smaller than random variations. So stop measuring it on a monthly basis.
I thought it was interesting seeing the cruft of earlier versions in the automatic update cycle (more noticeable in FF than chrome, about 2% total of market share) that doesn't seem to be disappearing. It would be interesting to see how that changes over the next, say, 5 years. To see if it decreases and finds a level that it stays at, or to see if it increases and becomes an issue.
The title is pure linkbait. The article doesn't make any kind of deal about that it it's such a small amount alot of sampling errors could account for that. The article tries to justify it's > 10 point difference from other surveys re: chrome vs IE so I don't think a sub 1 point gain is worht spending too much time talking about and the article doesn't.
If you go to http://marketshare.hitslink.com/, it also shows that Windows XP had an increase in marketshare. So my guess would be that some huge market (China?) added noise to the numbers, due to their methodology based on "estimated internet users per country".
This is what I was thinking. This is (hopefully!) not due to people choosing to use IE6, but likely due to people installing a likely pirated copy of XP who were not previously Internet users.
IE6 (and XP) still dominate in China and there are approximately 420 million Internet users. Over 90% of software is pirated, most Chinese users never connect to Windows Update or even upgrading their web browser. Also a lot of Chinese websites have been built to work with IE6 only (e.g. banks).
I suspect it's partly due to businesses continuing to use XP. I know of one company that's set-up a major division locally, and all their new computers are XP with IE6 - which is something to think about if you have a start-up that caters for business users.
When I read the title, a part of me was hoping that somewhere in the world it was still April 1st and this was a very distasteful April fools joke!
Interestingly enough, if you look at the "Internet explorer version adoption" graph, IE6 and IE7 appear to have identical adoption trends. Maybe this is some kind of sniffing error?
I think that's actually a shortcoming of that type of line graph. The lines move in parallel, but only IE6's adoption is changing. It's even worse for Firefox's graph -- it's really hard to determine the absolute adoption change for any given delta.
only if you're pulling from Windows Update. Many of the IE6 installs (large corporations) use WSUS, where they can pick and choose which updates to push to their users. I can see them rejecting the update if their business still needs IE6.
My company pushed out IE7 to all XP users, and that's as far as XP will get for us. The users still on XP (like me) have apps we use which will only support XP and only support up to IE7. Apps are being rewritten to get around this, and XP support will be dropped when that occurs. From talking with other companies at various Cons, many corporations are still working through this process. It goes slower if they don't have an application security team willing to push this every quarter.
I would wager there are very few home users who still have IE6.
Take a look at the graph[1]. IE6's segment is generally trending downwards, but the "wiggle" from month to month is a bit bigger than the trend. This month it happens to be wiggling up. Next month it will probably wiggle down, and ars will get to write a breathless article on the impending demise of IE6.
The correct take-away is that IE6 is declining, but slowly. So slowly, that on a monthly timescale, the change is smaller than random variations. So stop measuring it on a monthly basis.
[1] http://static.arstechnica.net/2012/04/02/internet-explorer-a...