Why Linux Doesn't Spread - the Curse of Being Free ("free = not valuable")

This is way off. Free always beats paid. Your example of champagne is somewhat accurate (i.e. http://www.timesonline.co.uk/tol/life_and_style/food_and_dri...), but I don't think the same concept applies to software. When some alternatives cost money, consumers will almost always choose free. Why do people use Windows?

(1) It is pre-installed on machines - note: Linux is no longer "free" when you need to install it yourself

(2) It is included in the PCs price or consumers aren't aware of or given "free" alternatives

(3) Historically, *nix has been too difficult to install/use, lack of software, etc

Think about this: how much did MySQL just sell for? How much does MySQL cost?

Hmm. You still have to consider the distributions of linux that are in fact very expensive. Look at Novell's distro. It's so fancy you can only rent it a few years out. That common man's XP you can keep pretty much forever. It will still be interesting to see the results of his experiment, though I think he will need more time to make a solid conclusion. One week doesn't seem like enough to me.

P.S. I know my argument about Novell ignores the fact that he was using pirated examples. I'm pointing out that even similarly priced OSes don't get as much market share as windows.

I'm not sure I buy into this argument. Most non-techy people I know have no idea how much a Windows license costs.

I'm often asked whether a specific offer for a PC "is a good deal" - invariably this is something off eBay which comes without an OS. When I mention how much a Windows license will cost on top of that, I'm usually accused of lying. (until I prove it)

Roughly as many times, I've been consulted when people "can't find Word" on their brand-new PC. For some reason everyone seems to think it comes free with Windows.

The most frequent piece of computer-related advice I get asked, though, is what "good, free virus scanner" I recommend.

I know the human mind is irrational and backwards, but is it really that backwards?

I really think it's mostly because Linux still requires a high degree of techyness, and the product and hardware support isn't great unless you really know what your doing.

It's still a pain to get flash or java on ubuntu, openoffice won't deal with docx formats, powerpoint isn't available, printers are a pain to set up, even modem support is complicated and bad out of the box -- I had to setup a virtualization layer just to get my wireless working.

It really shouldn't be any wonder it's not getting widespread adoption; there are confounding factors, but I don't think we need to posit price signaling to make sense of this.

This post reminded me of Kopelman's "Anti-penny gap" http://redeye.firstround.com/2008/01/some-thoughts-o.html

Price signaling is very real. High price almost always signals quality. Just look at luxury industry and how they can sell $30 worth of goods & labor for $1000s of dollars.

The main hurdle I experience on a daily basis is not the perception of value but the fact that being free creates mindshare battles that work against distributed collaboration and contribution. It is MUCH more difficult to organize collaboration around content that can be freely replicated.

Why? Users tend to migrate towards interfaces that focus on content presentation rather than those which focus on getting stuff done. This works in favour of WYSIWYG style projects like Wikipedia but works against OSS-driven web services.

One solution would be for major search engines like Google to skew rankings based on the presence or absense of generic AdSense-style advertising. This is unlikely to happen as long as search functionality is tightly knitted to ad networks.