Oxford says Apple's OS X is a Linux distro 9296

Actually, they are meaningful as long as:

A) you are using a broad enough sample that encompbuttes many clbutt of user. B) You are counting visits, not hits. C) You understand that this is telling you something about Internet connected users, not all PC users.

They are certainly more meaningful than sales figures, which fail to take into account free Linux downloands, redployment of old hardware, upgrades vs new installs, channel stuffing by vendors, and many other factors that muddy the water.

That being said, useful browser stats are hard to come by. Google no longer publishes their info, and they have arguably one of the broadest samples we could work with. Their are a few marketing companies that accomulate stats from a variety of sites and generate reports based on it, but the samples are smaller and the raw data unavailable for peer review without shelling out big dollars, so their conclusions are hardly useful.

For example, a recent OneStat survey put Linux at 0.26 percent market share, yet several years ago Google showed Linux at a solid one percent (even when not adjusted for spoofed or unknown browsers). Linux has continued to see year over year deployment growth of at least 20 percent since then, so a 75 percent drop in Linux is not possible.

The more likely reason? OneStat is working with a flawed sample, perhaps examining web sites that attract mostly Windows users or IIS servers that actively discourage Linux browsers. This could be especially possible if the study was sponsored by a particular vendor that might benefit from a particular finding. Alas, we might never know, as the details and methodology of their survey are not available as far as I can tell.

Thad