The title of this post is a quote from sci-fi scribe and Infoquake author David Louis Edelman. he elborates:
So why am I down on open source's prospects for the long term? A few reasons:
1. Open source hasn't proven it can produce better products, just comparable ones.
The reality of the matter is that most open source software is not
entirely ready for prime time. It's buggy, its usability is generally
wretched, and the documentation tends to be rather slapdash and hard tofollow. In short: pretty much like regular, proprietary software.
2. Software is too cheap to worry about saving money on it.
Windows costs $150 or so out of the box; less if you upgrade or buy it
pre-installed on a new computer. And it comes with almost everything
the general user needs: basic word processing, web browsing, a mediaplayer, e-mail. So if I'm a business owner, the question is, why shouldn't I pay $100 per user for a product with complete user documentation and technical support?
3. As software gets more complicated, open source will have a problem keeping up.
We've got limited space here on Earth for people, and only a limited
percentage of the population qualified to write software. But with
quantum computing on the horizon, there's nearly unlimited potential
for software and no reason that programs won't get more and more
complicated. Do the math: eventually there will be far too
much code out there for us to have a population of open source experts
on every piece of it. And when it becomes a million times easier to
exploit software than to fix it, the first obvious precaution is to
lock up access to all those exploits.
Now, Edelman is WAY smarter than I am, and he wrote this in defense of his own futuristic book setting (which is to say he's published, and I'm not), but I think he's overreached a bit here. I don't think open source software is going anywhere, and by that, I mean that I don't think open source software is going to get any more popular or pervasive than it is now, but I don't think it's going to get wiped out, either.
All the reasons Edelman lists above explain why the individual consumer and even the small to medium sized business (SMB) has no reason to use OSS. All the financial gains of "not paying" are more than lost in convenience cost. The average user doesn't enjoy or have time to spend dinking around with the infiintely customizable quirkiness of Linux. (Incidentally, Apple is on the other side of this equation, being SO specialized but so non-competitively priced that its financial costs outsrip the convenience gains. In essence, Apple is a luxury brand.) Mainstream, proprietary software—which is to say Windows—hits the sweet spot of being reasonably reliable and reasonably affordable in a one-off setting. Not perfect, but good enough.
This breaks down in a bulk-usage setting, like server farms. Here, small per-unit cost gains add up quickly, and since you're dealing with large-scale technology, you're already in for having to pay for quality human-side expertise to protect your investment. Thus, the convenience cost of open source software is already part of your overhead, so why not go ahead and save on the financial cost? That's why Linux is all over the bulk server market, and why OSS is not likely to disappear form this area. There's also academia, which is generally burdened with a lack of funds and an overabundance of free-to-cheap semi-expert grad student labor, which also explains why OSS flourishes there. (Plus, there's something to be said for the share-and-get-street-cred nature of academia versus the lock-up-and-outperform realities of the market that make Linux the preferred OS of the ivory tower set.)
On the complexity front, Edelman has a HUGE point, but there's also something of a law of diminishing returns, which suggests that at some point additional complexity merits you nothing. Install Win95 on a Vista-ready PC, and watch the old girl fly. Is there really anything in significantly better in Vista or Word 2003 that wasn't available in the '95 versions? If anything, increasing software complexity will spell an end to the arbitrary upgrade cycle that Micorosft and Intel have everyone chasing. Most folks only upgrade the software when the hardware dies, anyway. Imagine a world where developers are competing to write leaner, meaner, more efficient software, rather than force us to buy expensive new processors to tweak a library of clipart and fonts we already don't use a tenth of. Now tehre's a sci-fi setting I can get behind.