I interviewed Bernard Golden, CEO of Navica, the other day at the 2006 MySQL User’s Conference. Bernard is always worth listening to, and hopefully our audio interview will appear on Distributing the Future sometime soon.
I hope it’s fair to say that Bernard believes that the economic advantage of open source software is so compelling that companies cannot avoid it.
That’s curious, if you think about it. If it’s difficult to argue that not only do people distribute high-quality software for free and with generous redistribution policies, but that that software can be more worth using than proprietary software, there’s an assumption or belief mismatch somewhere.
Consider the normal economics of value. Scarce items (with some degree of value) cost more than plentiful items.
As open source advocates know well, software is cheap and nearly free to copy. The costs of creating the software and the costs of equipment and resources to use the software are huge in comparison to the cost of making a copy of a piece of software to run. It’s always been that way.
Why is software so expensive? Where is the scarcity that raised prices beyond the nominal fee of copying? Some people argue, legitimately, that the unit price for a proprietary program amortizes the more expensive costs for producing the software.
To some degree that’s true (and to some degree it’s false; why has the price of Windows XP not dropped over time?).
An obvious open source argument is that amortizing the cost to develop and polish software over multiple people not tied to a single company or economic organization reduces that cost as well. If you consider the available pool of software developers and software participants as a commons, the cost of developing any single piece of software tends toward zero over time.
(Now it’s true that certain software is easier at certain points than others. Xlib begets Motif and Qt and GTK and C begets Perl and Python and Ruby and hard problems get easier year after year. You can assume that everyone has a web browser and can install a web server and a database now; try that in 1994!)
Removing the artificial scarcity of copying software and spreading the cost of creating that software across multiple projects and a much larger potential pool of contributors should drive the cost of software down — and it’s clear that where open source works, it does.
Here’s my question.
Where does that value go? Does it leave the economy? Or does it go elsewhere?
It’s not as silly a question as it seems. Removing artificial scarcities from the economic equation should make the points of true, unremovable value more clear. Knowing what’s really valuable should help programmers and other people who want to make a living from and related to software somehow compete and even thrive.
Part of that value and part of the money that companies would otherwise “invest” in companies exploiting the artificial scarcity of copies of software will likely go to companies and individuals charging the real cost of actual products of value.
I think part of that capital will go to new things.
Would Google be possible without Apache or MySQL or Linux or the web or DNS or TCP/IP? Could two graduate students build a 5000-person multi-billion-dollar company for the artificially inflated costs of all of the software they use? For every person making money off of Google stock or displaying and selling text ads on websites or map mashups or this or that, consider that this source of revenue is the result of years of investment by a huge commons of people who probably never anticipated such a result.
Open source is the printing press. It’s always the printing press. If you see yourself as a scribe, that’s scary. If you like the result — ubiquitous information — that’s wonderful.
I don't think, at least at this point in time, that open source shifts value to items other than software but only further up the software development stack. Open source components just mean that Google neither had to develop nor buy OS and web server and database software...good thing, too, considering how much it clusters. But Google still had to develop cutting-edge web search software that used those components. The company still invested value in software, but just at a higher level. Its revenue is also an interesting case because its main product, advertising, is deeply dependent on its popularity, an "apparent" value. Its inflated stock price is the same. Google is in the unique position of providing value to the economy because everyone says it does. Of course, I guess the number of web searches done on Google, the number of pages it indexes, etc., objectively demonstrate that it has prime value for advertising.
On the other hand, there are complete, open-source, end-user products like a CMS that a business can put in place for its software without developing any new software at any place in the stack, but even then a lot of open source software is not sufficiently easy to install and use as the proprietary counterpart would have been. That means in addition to the zero cost of the actual bytes the business must absorb the cost of getting the software up and running smoothly and then keeping it going.
My point is that although open source is distributing the cost and value of software among large numbers of people (you incur a small cost to make the software a little better but you still get to benefit from the full value of the rest of it), there's still a ways to go before software ceases to have an economic cost of some kind associated with it. In terms of ubiqiuitous information, open source is even better than a printing press because the cost of copies is lower.