The latest news that’s got everyone aflutter is a report by Mobile Application Data reporting company Crittercism compiled between December 1st and 15th last year relating to application crashes on a spread of various iOS and Android versions. The data actually seems surreal (unbelievable even!). The analysis examined app crashes at launch, and is further broken down into normalized data split between iOS and Android versions. Now, remember this is a percentage of apps that crashed, not the number of apps. The data also showed that more apps crashed on iOS5 than on other versions of iOS (which should be kind of a given since apps have to be updated by the developers to support iOS5, which usually happens on time, and by the user on their device, which I can say happens less frequently).
The chart on the left (on top) breaks down app crashes during launch by quartile (or percentile if you will). The first quartile represents 25th percentile, second quartile is the 50th percentile and the third quartile is the 75th percentile. So what does this mean? That the outliers are the first and third bars in the bar chart. The chart on the right is the normalized breakdown of app crashes by OS version. Apart from iOS 5, there’s a large portion of iOS app breakdowns that can be attributed to iOS 4.2.1 (people either haven’t updated their devices, OR, the more plausible answer, to me at least, is that these are the people who have jailbroken iOS devices that haven’t updated their OS, for various reasons). The highest percentage of crashes on Android are on 2.3.3 and 2.3.4. Android 4.0.1/2/3 doesn’t really figure in the data much, which isn’t that surprising either; it’s new and is on barely 1% of Android devices, officially or otherwise.
There’s more granular breakdown for each OS; Android and iOS, in the charts above. Again, the breakdown by OS version is pretty self explanatory.
Now, I’m as surprised as everyone else is. I’m sure everyone else is too.
You can read a more granular breakdown of the data (with a more detailed explanation for data buffs) over at Forbes.