Bruce Perens: Innovation Goes Public


Bruce Perens speaks to
UOSC
Bruce Perens speaks to UOSC
(click to enlarge)

Bruce Perens is here for the second keynote of the evening. I spoke to Bruce on IT Conversations last March when he was last in Utah to protest the Novell/Microsoft deal. Bruce is not only interesting to listen to, but entertaining as well.

Bruce gives an intro about why he got excited about open source. As I mentioned, Bruce has criticized Novell in the past (and will today). The conference is being held at Novell, but the security folks haven't thrown him out yet.

How can "innovation" go public? It's not a company. FOSS development share a lot in common with investment markets. Note: this isn't about open source companies going public. It' about a new way of innovating in a public setting.

Open source isn't Redhat, Novell, or HP. They're part of the community, but not the majority. Open source is development teams that make products. Apache, wikipedia, and others are products, not companies.

Open source appears to go against capitalistic economic theory, but only until you look closely. It's main developers are not it's vendors. It's very customer -centric in its emphasis, rather then vendor-centric. (Cue Doc Searls here.)

Companies go on the stock market to distribute cost and risk with the motivation of sharing in the reward. Non-open source software developers must pay for development first and then start making a profit. The stock will be worth nothing if the company fails. Stocks allow many people to share in the cost and risk with the hope of a reward.

FOSS distributes the cost and risk of development among many developers because each share sin the development and no one puts in too much. The pay-back is working software that is high quality and well suited to it's user's needs--because some of them wrote it. FOSS developers float their source code on the market for others to share.

Amazon, Gogle, Merrill Lynch, and Pixar are all companies that give away software and still make profits. They do so without giving away their critical innovations.

Most companies don't sell software--they sell books, wine, services, and so on. But all of these companies still need software. Software isn't a top-line product for most companies, it's a cost of doing business. Most of the software that companies need is non-differentiating. They can give this away and it doesn't make any difference in how customers perceive the company. Giving your competitor software makes them your partner--in that non-differentiating area. Bruce guesses that about 95% of the software in any company is non-differentiating.

Companies should naturally try to remove as much of their budget from non-differentiating software and move it to the innovative software in their business that differentiates them. Of course, that means that you have to find a way to get the other 95% of your software as cheaply as possible: enter FOSS.

Bruce points to HP and IBM. HP had plans at on time to spend $1 billion rewriting HPUX. Maybe they still are. IBM, on the other hand, decided that AIX wasn't a long term differentiator.

How do you find differentiating software? If yur competitors can get it, it's not differentiating. So, neither Microsoft not FOSS is differentiating. For software to being differentiating, you have to control it. So, your employees or consultants (as long as you control it) have to create it for it to be differentiating.

Business differentiating innovation has high value to a company. Non-differentiating innovation improves some internal aspect of the business, but doesn't make a difference to the customer.

Sharing non-differentiating innovation between companies has a low transaction cost. Open source has always trumped consortia (think X.org vs. X Consortium, Linux vs. Taligent, and GNOME vs. CDE). Open source structures put product first. Consortia put vendors first and they work to the detriment of the product.

People with similar (crazy) ideas can form on the 'Net and create real product with no capitalization except the sweat of their brow.

Open source is a massively parallel drunkard's walk filtered by a Darwinistic process. The result is that 10,000 people all do what they feel like and the result is Linux or Apache. Open source isn't like a company, it's like an ecosystem. Projects start with one person's idea, but they don't become open source until they're at least a little useful to someone else. Lots of projects die before they reach critical mass. They don't waste much resource.

Strategic marketing, on the other hand depends on someone predicting the future. Being smart isn't good enough--you have to be prescient.

When an open source project dies, someone else can take over--anyone has the right. This strongly protects companies that use FOSS from having the software become useless. Source code escrow doesn't work: judges frequently nullify them.

Should FOSS be allowed to displace proprietary software? Should refrigerators have been allowed to displace the large ice industry at the turn of the century? FOSS should be allowed to displace proprietary software when there are clear advantages. Of course, that's a political statement.

If you can take FOSS into a business, take the low-hanging fruit rather than attacking Microsoft on the desktop. FOSS is being used in almost every business and become a mission critical function--often without the business executives being aware of it.