carpe decorum

Various and Sundry

In Debunking the Linux-Windows Market Share Myth Nicolas Petreley makes several important points:

1. All PCs with WIndows OS installed are counted, even if they are sitting as unsold inventory. An upgrade from one version to another appears to add to Windows market share, and not be counted as an offset.
2. Linux OS market share is calculated based on downloads and sales, and does not include inventory. Downloading several Linux distributions, but choosing only one is not captured in the market share.
3. Linux runs on non-PC platforms and tends to outperform Windows on Intel-based machines; by implication fewer Linux boxes are needed to run the same tasks as Windows servers, skewing market share ostensibly in WIndows' favor.
4. If a Windows OS is replaced by Linux, WIndows market share does not drop by one.
5. Tracking OS on machines is notoriously difficult at large corporations.
6. Developers appear to prefer Linux to Windows.