MONEY MATTERS

Dick Pountain/13:08/04 March 1996/idealog 19

Should all software be free? There's no denying that this is a serious question, if only because there are lots of people who believe the proper answer is yes. In fact when you examine the subject closely it turns out to be a rather difficult question indeed, and one with wide reaching implications.

I think we can take it for granted that Bill Gates would be a no, though I'll confess that I didn't bother to ask him. Equally certain is that Richard Stallman - legendary author of EMACS and GNU and president of the Free Software Foundation - would be a yes, as would all those thousands of Unix programmers who subscribe to the FSF philosophy and donate their time to writing free software tools. In between lie the magazine reading masses (that's right, you) who appear to have an insatiable appetite for cover CDs crammed with free software, but probably rarely consider the implications.

A free-market economist would argue that software is a product, and that by giving it away free you fail to collect any revenues that can be reinvested to produce more product, and so eventually there will be no more software. Furthermore, by underpricing the product you will create an enormous demand for it, according to the laws of supply and demand, and that should lead to shortage. A Marxist economist (now there's unfashionable) would add that the labour cost of producing software must at least cover the cost of reproducing the producers; in other words the community of freeware programmers presumably have to eat like other people (well maybe not EXACTLY like other people) and therefore require a wage.

In fact many, if not most, of the people who write free software are professional programmers who do charge for most of their time, so their commercial activities are subsidising their freeware activities. For that not to be the case you'd have to build a complete alternative economy that ran by bartering software, but even in California there are few establishments that will give you a hamburger in exchange for 25 lines of C++.  

However it's the supply-side of the economists argument that intrigues me. What is so novel and disconcerting about software is that it costs effectively nothing to reproduce (pedants may wish to work out how much electricity is consumed in copying each byte from one memory location to another) and so the predicted shortages just don't happen as you just crank out another copy. We are quite familiar with the concept that a product need not be material - after all copyright law and intellectual property rights have been in force for centuries - but there was always a significant reproduction cost that helped to enforcing these rights. Books had to be printed, and even though the invention of the Xerox machine lowered this barrier it didn't eliminate it, so few people (outside of the old USSR) would Xerox a paperback rather than buy it. But software is so cheap to reproduce that it seems to just flow where it will, like air. Which is why some cypherpunks and net-libertarians reach the conclusion that information wants to be free, and that the doom of capitalism is nigh.

However what is definitely not free is the hardware that software runs on, nor the communications bandwidth that is required to convey it from one computer to another, and both of these resources fully comply with the laws of economics. Computer hardware companies that get their pricing policy wrong can go bust, be they as big as IBM or Apple, while we won't get fibre networks running into our homes until someone decides who will pay for it. What's more the congestion that you now encounter every day on the Web testifies to the truth of the economist's argument - the Internet is underpriced so too many people use it, and hence create a shortage of bandwidth - and the arrival of Internet phone programs like WebTalk can only make things worse. Last week it was announced that some new portions of the Internet will in future be reserved for the scientists for whom the Net was originally constructed.

The free software people have decided to treat their work in the same way that scientists do, which is no coincidence since it's mostly Unix software and Unix is the OS of choice among computer scientists. In the world of science results have always been distributed freely between practitioners, and indeed science cannot be properly conducted if the free flow of information is hindered by patents and commercial secrecy. As a consequence scientific activity has to be paid for as a social good, ultimately by the taxpayer, rather than by the direct sale of its products. However it seems extremely unlikely that in a political climate where even the financing of basic welfare is being called into question, you could ever persuade the populace that software production ought to be publicly financed. 

Whenever I enter a W.H.Smiths and see half a dozen computer magazines with CD-ROMs on their cover each containing over 100 Mbytes of software, I can't help wondering for how long these scores of gigabytes of free software can be dredged up.

----------------BOX OUT [WITH KICKING COMPUTER ICON]-------------------------

CYBERPHOBE OF THE DECADE
A story in yesterday's financial pages prompts me to revive this occasional feature. It seems that thanks to a shift in stock prices, financier Warren Buffet has just overtaken Bill Gates to become the USA's richest man. Plain-talkin' Buffet still prefers to dine on hamburgers and hang out with the good ol' boys, and it seems that the Omaha headquarters of his Berkshire Hathaway company doesn't employ any computers.

----------------END BOX-------------------------------------------------------