Yes, its very much a matter of Intel not having its drivers for Sandy Bridge ready when the chip was released rather than Intel drivers being bad in general. I got a new Nehalem laptop right after they came out and the drivers barely worked at all for a couple of months. Then there was a new kernel and everything worked beautifully with no problems up until the present day. Is it bad that Intel's Linux drivers are months behind the release of new hardware? Certainly! But for older stuff they work great.

"Buy old hardware" is not a way to get people to stay on or move to your platform of choice. People don't care why their stuff doesn't work--just that it does not.

A self proclaimed linux enthusiast like the ranter is well able to determine which hardware will and won't work before he purchases it. The fact that he didn't care to do that and then complained bitterly about it suggests that's not really the problem. He could have spent less time than it took to write the rant on the front end to end up happy.

All hardware just works isn't a quality linux has ever suggested applies to itself. But that is hardly rare for an OS. The only OS that applies to is Windows (generally) and that's because that's the only platform all vendors bother to target. You can't buy arbitrary hardware and expect it to work at all under OS X, Free BSD, Solaris, Android Open Source Project, Windows Mobile and iOS.

In fact, out of that whole list the odds are that an arbitrary hardware configuration will work out of the box are highest on linux.

If you want any of those platforms to work well with a new purchase you have to determine beforehand the compatibility. So linux is hardly some sort of freak in that way.

But don't let that get in the way of a good rant.

You're assuming people care to evaluate Linux in the way Linux advocates want to evaluate it. They do not.

"Hardware just works" is a characteristic that normal people consider important to an operating system. If you don't do it, you are deficient in that area. "But all these other also-rans fail at it too" is not an excuse. The standard, for better or for worse, is Windows, and you either meet the standard or you don't complain that people get tired of dealing with your failures and move on.

And it's certainly not surprising that even a Linux enthusiast gets tired of dealing with substandard, broken hardware after a while. I know, I was one. Eventually I got tired of trying to balance having decent, modern hardware with having hardware that supports an operating system that, on the desktop, is essentially a novelty (it affords me nothing that Windows or OS X doesn't), and I tubed it. Because it wasn't worth the effort.

For most people I'd say the answer if to find a machine with Linux pre-installed.

And pay more money for less quality? (It's not as if pre-installed Linux is cheaper than pre-installed Windows, and manufacturers that offer it aren't exactly in the highest of tiers.)

The only way Linux makes sense is if it is competitive on at least one level, and it currently is not.

I was recently in the market for a laptop that could be ordered with 16GB of RAM and two hard drives. The options I found were the the Dell M6600, Lenovo W520, and System76 Gazelle. System76 had the greatest range of options available and the lowest overall price. The best part was that the options were priced near their cost; going from 4GB to 16GB of RAM added only $280 to the cost compared to $480 from IBM and $800 from Dell.

I haven't made detailed comparisons, but I expect System76 would be competitive for more mainstream laptop configurations as well.

Quality doesn't mean only specs. System76 isn't competitive in terms of after-purchase support (Dell or Lenovo) or in terms of build quality (Lenovo).

The reason people pay more for a Lenovo machine is because it's built like a tank. You pay more because you get more. System76, on the other hand? You pay less and you get a lot less.

EDIT: System76's build quality looks even worse than it used to--those are truly hideous plasticky machines. They're also pretty underspecced and crappy, on the low end. The only processor available in their 13.1" model is an i3-330UM running at 1.2GHz. My x201 has an i5, and you can get them with i7's. It skimps in the peripherals, too: it comes with a 5400RPM hard disk and a smallish battery. When you up it to a 320GB 7200RPM disk and a halfway decent battery, you're already up to what I paid for my X201. Adding 8GB of RAM puts it over (although that's somewhat unfair because I bought my RAM off Newegg).

The machine also weighs more than my x201, has one of those weird button-y keyboards, no trackpoint, has a glossy screen, and less pixel density. Fails across the board. So, at least for me, System76 doesn't even begin to compete with Lenovo. Even if you don't consider Linux on the desktop a minus, it doesn't make a lot of sense to buy what looks in most ways to be a cheapie Dell knockoff.

Clarification: where I refer to my x201 having an i5, that is to say it has a non-ULV i5, running at 2.4GHz.

There's an inherent problem with releases and the range of bug-tolerance in the Linux community. It's hard to know when something is "ready," unless you are a bleeding edge nightly builds kind of person.

For example, KDE4 caught a lot of flak for shoddiness and flakiness before its first release. That happens to a lot of software, and it's completely unfair. I criticized people for expecting KDE4 to be polished enough for a general audience before the point-0 release. However, when it actually was released, it was still in terrible shape. I first installed 4.1 and gave up on it pretty quickly. Not only was it woefully incomplete, but the aspects where 4.1 caused me the most grief (cosmetic glitches, unpolished UI elements, non-intuitive and non-discoverable customization) were supposed to be major focuses of the KDE4 effort.

Then it was my turn to be criticized: KDE 4.0 and 4.1 were intended for early adopters, and it wasn't until KDE 4.2 that KDE4 was deemed a good replacement for KDE3. Well, I admit I didn't read the release announcements, but what the hell is a major release supposed to mean, then? What does it mean to go from beta or RC to point-zero if the point-zero release isn't ready for most users? There's a rule of thumb that conservative users who absolutely need stability are wise to sit out an x.0 release, but there sure as hell isn't any convention that most users should wait for an x.2 release. And why did my distro suggest 4.1 as the default KDE to install?

Clearly the open-source world needs a more explicit and widely-followed convention for labeling releases. Projects are willing to cheat and release x.0 versions that aren't ready for general consumption, just so they can make an artificial deadline, keep up with their competition, or even just to ensure plenty of users for testing a new version. And they have no shame about it. If a project as high-profile as KDE does it, anyone can defend it based on their example. There needs to be an explicit convention for labeling releases so each user knows which release they should wait for.

x.0 releases often occur because it's the only way to get a sufficient number of testers. I remember reading this as the rationale for pushing KDE 4.0 when they did, and it's generally true in all software no matter how well funded the QA department is. You may experience a shoddier x.0 with OSS since most QA budgets are around $0, but even products released by big companies (OS X and Windows are two comparable products ;) ) have some serious problems when a new major version is released.

It seems to me that the IT industry has come upon dot-oh conservatism honestly. Isn't there a persistent meme to not ever roll out a dot-oh release or a microsoft OS until the first service pack? Plenty of people seemed to feel that way about deploying Vista.