Dividends and Free Software

People like me (that is, people who are paid to write computer programs, not people who compulsively rant on the internet) should probably be considering the meaning of free software. By free I don't mean as in freedom, but rather free as in beer and freedom. While crazi^H^H^H^H^H visionaries like Richard Stallman emphasize the difference, we cannot deny that one of the impacts of having free(dom) software is to drive the price of that software down. You can buy Linux (or rather buy a distribution, maybe with support), but if you would like Linux for free, that's a very, very viable solution. (For my three installations of Linux, I have paid a total of $0.)

For the sake of argument, let's call this phenomenon "cheap" software - that is, as a user, your total outlay of cash for software might be a lot less than it used to be as (1) features you used to have to buy become integrated into components you own (e.g. compression/decompression is now free with every major operating system) and (2) software that you might have to buy has become free in itself (e.g. the operating system, your word proessor).

Why is there free software now (when it wasn't such a major part of the software ecosystem 20 years ago)? I think there are perhaps three reasons:

1. Communication technology has improved, making it possible to aggregate programmer effort easily. (In other words, we can all work on code over the Internet.)
2. Dividends from faster hardware have been invested in programmer productivity. (In other words, we get more done because we're not all coding in assembly language.)
3. We have free time.

I think there is no question that having the price of software decline is a good thing. It would be absurd to argue that improvements in farming are bad because it means we have fewer farmers making most of the food - we can all recognize that without this, no one could take jobs making cars and houses and everything else that makes us, as a society, wealthy. In fact it's this very wealth effect (that is, increases in productivity mean we have to work less) that makes free software possible.

What got me thinking about this lately is my friend pointing out that Microsoft is offering people cash to use Windows Live Search. I see this as a sign that Microsoft is really, really screwed. Thy are a huge company, but their glory days are over - they will never again control a monopoly like the one they had on the desktop. Pundits have argued that they need to transition into a large, stable, diversified company, but I don't think it's going to happen for two reasons:

1. Computer technologies tend to be winner-take-all. If Microsoft can't take all of a new hot area, they won't be the winner.
2. The price of software, their core business, slowly and perpetually moves to $0.

If Search is the next big thing, it's really too late for them to try to claw their way back, even with cash. The winner (Google) is already taking all.

Remember the good old days when a PC cost over $1000 and it was a good business to make PCs? What happened? We hit a point where consumers didn't want more, they wanted cheaper. Rather than put the hottest new chips in the computer, manufactures had to compete on price, and now you can get a PC for $400. It's not good to be in that market any more.

The operating system has gone the same way. Users don't want any more operating system. Now that every major operating system has protected virtual memory, tolerable parallel performance, a journaled file system and a GUI on top, there just aren't new interesting features. I don't know what will be new in OS X 10.6 and (at risk of blasphemy) I don't really care...there's just nothing more Apple can add to the OS that is interesting to the vast majority of its customers. Whatever is in the next Windows, the only thing they can do right is clean up the disaster that is Vista, and that hardly makes a hit. (You shouldn't have to release a broken version of your software just to create the opportunity to improve in the next version.)

In this environment, what users will want is not new features, what the will want is something that works and costs as little as possible. Where could I get an operating system like that?

It will be interesting to see how the battle for embedded platforms plays out. The difference betewen the desktop and the embedded markets are:

• Open source is already here while the market develops - no head start for commercial software companies.
• The margins are very, very low - price matters so much more.
• There is no legacy of X11/Motif - free software isn't starting with a ball and chain tied to its ankle.

I'm not betting on Microsoft.