Is There any Hope For Desktop Linux?

@notbob: Do you not realize that marketshare also correlates to the number of volunteers available to write drivers (i.e. the volunteers are users of that OS)? Secondly, the BSDs and other really obscure OSes like BeOS (I'm sure there's people still using it) have puny marketshare, and they're famous for not having any drivers. I don't see any single fellow writing 160 drivers for those OSes.

If you're a programmer using an obscure OS, you're either going to wait for that company to give you the documentation you need, or you're going to reverse-engineer a driver on another system. The second route is very tedious and time-consuming, and isn't used that much. You sound like someone who isn't a programmer or engineer at all.

Yes, Linux had better compatibility than W7, largely because it supports a lot of older hardware that W7 doesn't. But also because it has lots of people working on it (== marketshare). Contrast that to OpenBSD; it has terrible hardware support. Where's the hordes of volunteers writing drivers for OpenBSD? It's simple: OpenBSD has much lower marketshare than Linux. Lower marketshare = fewer users = fewer people willing to develop drivers.

BTW, system administrators generally don't have CompSci degrees. And yes, they do teach Windows at many Universities. Frequently, they teach Java on Windows.

You don't need to register as a Linux user to be counted in web stats, you just have to visit a website with your browser running on Linux, and not change the user-agent string. That's why everyone looks at web stats for this information these days, because obviously sales numbers are bogus.

@smallboxadmin: Who cares? As long as it generates good conversation, the initial article really doesn't matter. And like Koriel, I didn't read the article either.

@JaseP: I'm not sure what the exact definition of "marketshare" is, but it seems to me that what's more important is how many unique Linux users there are, not how many unique Linux computers there are, although I guess if you're looking at it from a HW maker's POV, they might care more about the number of computers since someone who buys more hardware is worth more to them than someone who buys less HW (though of course this might very well depend on the exact type of hardware.

However, I have to disagree that any FOSS is worthy of support. Companies have limited resources, some more than others. Not all of them are giant behemoths like Intel and MS. I work for a hardware company that has all of about 35 people. If someone wanted us to write a driver for them to run our hardware on, say, FreeBSD, we'd have to tell them no, because we don't have the resources to put together a new team to write a driver for a single customer that won't be buying very many units. Saying companies should spend far more on support than they're going to get back in sales is just inane. That's why marketshare is important: if we can show that there really are a certain number of Linux users out there, HW companies will see that they need to support it to improve sales. After all, telling 1 in 20 of your potential customers to take a hike isn't a good idea. But if they think that it only has the marketshare of Minix, they're not going to bother.

@Icafiero: It doesn't need to be a market share leader on the desktop, but it needs to have something that's viewed as significant, or else it gets zero support from hardware manufacturers. Ask the BSD fans how it feels to not be able to use all kinds of different hardware that we Linux users take for granted. I can plug in many wacky kinds of USB devices to my Linux box and they "just work". I can buy just about any motherboard or laptop and things mostly "just work". I don't have to worry much about some southbridge or audio chip not working. I'm pretty sure the BSD fans don't have that luxury. The only reason we have this is because we have sufficient marketshare and mindshare to accomplish this. Many drivers aren't even made by the manufacturers, they're made by volunteers, but those volunteers are part of that marketshare; zero marketshare means zero volunteers, as the BSD fans have found out the hard way. But many times the volunteers are able to get specs from the mfgrs, and the marketshare stats help convince them to release this information. The higher Linux marketshare is, the better support we'll get. Sure, Linux is doing great in servers, but that doesn't help me much if I'm trying to run it on desktop hardware, which is very different and much more diverse than server hardware (you don't use USB dongles on servers, nor do you use laptops as servers). Linux doesn't need to be the market leader by any stretch, it just needs enough marketshare to not be casually ignored. A mere 10% would be more than enough, and an uncontested 5% would be good.

The Porsche analogy is highly flawed, since Porsche doesn't give their products away for free (they cost a LOT), and doesn't rely on free support from suppliers (they pay for their suppliers to support them, using the money they get from their high product prices).