This in turn makes it harder to track precisely how many users that an operating system like Linux has, but thanks to the web and the browsers that gladly indicate to sites what version they are and the operating system they're running on has allowed for at least some evidence as to just how many people are running their PC's with penguin power. However, these figures can be skewed by:
- Which sites are monitored and during which periods
- The fact that not all devices that may use Linux are used to browse the web, or are used less commonly (smartphones for example)
- Devices behind routers, especially as the amount of internet capable devices increases
It may also hide other important trends due to the fact that these statistics tell us little about the particular device that is being used unless the device requires a more significantly modified operating system that may be indicated in what it reports back to the sites the browser visits.
However, there are things I think we can safely assume about Linux uptake. People will be highly unlikely to simply switch when Windows or Mac is already working well for them due to the barriers involved (Installing an operating system, learning new software, moving across files and checking compatibility). This in turn requires opportunity and reason for someone to want to make the change.
An opportunity is best presented when someone feels the need to buy a new PC, with which comes the possibility to sell a computer pre-shipped with Linux where the user already has to overcome certain barriers anyway regarding typically laborious tasks like software re-installation, and may become a situation where Linux shines like package management systems which make the system easier to setup to how you want it to be. Other things regarding the likes file hardware compatibility are more easily taken care of by the hardware manufacturer too, like Dell who pre-installs common codecs and the ability for the hardware manufacturer to choose components that work well with Linux, alongside what may be an ability to strong arm partners into producing a driver necessary to use a piece of hardware, or at least make the process easier.
Another immense (and likely bigger) opportunity is new device categories. Entirely new, never before seen categories of devices present a new and intentional break from the past where developers and vendors won't necessarily encounter the same legacy issues in already established markets like desktops and laptops, creating a great opportunity for Linux to get a foot hold. Due to the new device category, there's less need to pander to already existing software that a user may need or want to use other than compatibility due to a likely required difference in UI and perhaps other user facing ways that software is developed and of course the use case of the hardware itself. This doesn't mean a vendor shaping Linux and appropriate software can do just what they want, just simply that they're not tied to the same familiarity and established high software demands that they would trying to compete in already existing markets, especially when they're happy enough to have the device that fits a need not met before.
The fact is, web stats are not a great measure of the specifics of trends like particular devices, and is really only best when you have no other alternative measurement that can give you those specifics like sales can. The only way sales can increase is if Linux software vendors properly capitalise on opportunities that may allow them to get to the larger consumer market, rather than myopically focusing on large businesses. After all of this babbling, it seems to me the biggest barrier to Linux adoption isn't a lack of ability for Linux to adapt or be easy to use, nor network effects as these can be gotten around. The biggest barrier is in fact the vendors themselves, who regularly are purely software companies that focus on support, which often causes them to focus on the more lucrative business market to the detriment of ordinary consumers. For any Linux vendor to be properly motivated to pursue people outside of the 9 to 5 schedule, they need to be in the business of selling to them, and that seems to be only possible if you're an integrated hardware and software company.