I've run a quick analysis of the Question Writer website logs to try to determine the prevelance and progress of the different .NET runtimes. I've also added some figures published on the Business of Software forum in March 08 by SteG for comparison.
I thought I'd publish this data as there don't seem to be any official numbers available, and there's a real dearth of information on it. (Compare this to the stats for the Flash player made available by Adobe)
Version 3.5 has been making great strides recently.
Right now .NET 3.5 covers 52%, .NET 3.0 covers 59%, .NET 2.0 covers 70%, .NET 1.1/1.0 covers 78%
Possible sources of error:
I'm using the same methodology as describe by SteG in the Bos post mentioned above.
Only Internet Explorer reliably provides .NET version information -
IE users use MS software and so may be more likely to have .NET installed.
IE users may be generally less inclined to install new software, and may be less likely to have .NET installed.
I'm identifying IE users by looking for the string 'MSIE' in the logs - this may be catching other clients which don't report their .NET versions. This may cause the 'Dot Nothing' figure to be artificially high.
Question Writer uses the .NET runtime. Visitors to the website may be more likely to have .NET installed. Particulary version 1.1
October data only covers first 12 days of October - it might be a little less reliable than the previous months.
There were some users of .NET 4.0 but there were so few, I've bundled them in with version 3.5
Note: I'll update the data if I get any good recomendations for improvement in methodology.
The results differ quite a lot from the April 09 data published here. I don't have an explanation for the discrepancies.