Comment by pona-a
2 years ago
I find this sort of metric rather pointless...
For one, how do we define Linux desktop? ChromeOS is a distribution of Linux, unlike say Android even advertising official support for a Debian container to run Linux desktop software, and while it lacks GNU utils, Alpine does too and yet we still include it in this category...
Next is the whole way we're counting the supposed number of Linux computers: user agents. There's zero reason for browsers to provide true and specific user agents, in fact, true user agents often yield broken sites (like YouTube assuming ARM PCs are all a specific Chinese smart TV) or pose a great privacy risk. Certainly a significant percentage of Linux users made the choice against it, either by themselves or as a browser default.
Then there's the fact specific demographics have their own distribution of platforms. For example, StackOverflow developer survey showed 40.23% of developers prefer Linux for personal use, compared to 31.07% who prefer MacOS. The absolute numbers matter little if your users in particular tend to be on a certain platform.
>I find this sort of metric rather pointless...
As a complete outsider, it might seem odd.
But I think tech people recognize this as referring to "OS's that aren't Windows or Mac".
> But I think tech people recognize this as referring to "OS's that aren't Windows or Mac".
Do they? The metric includes "Unknown" and "Other" not referred to in this, representing another 5% of activity.
However, what the OP seems to be saying is that interpreting anything from the metric and trying to communicate that interpretation with others is pointless. Not even the linked article can decide whether Linux desktop penetration is 4% or 6%, and that's just one person. Good luck when you add a second person with their own feelings.