Guns, Germs, and Open Source: Yali’s Question for the Software Business

by Dan Woods

I just finished Guns, Germs and Steel by Jared Diamond, and like pretty much everyone else who has read that book, I was awestruck and inspired. Like a couple of other authors before me (see end of blog for list), I’m seeking to apply Diamond’s analysis to the question of which type of software will prevail: open source or commercial software.

The book is a mind-expanding journey through Diamond’s analysis of why certain societies in the world dominated others. “Guns, Germs, and Steel” refers to the proximate causes of the domination. The book, however, is a detailed analysis of why certain societies developed “Guns, Germs, and Steel” that enabled them to prevail, while others didn’t. The book is framed as a long answer to a question that Yali, a tribal chief in New Guinea, asked Jared Diamond about why Europeans had all the “cargo”, Yali’s term for technology. (If you haven’t read the book, a rapid-fire summary has been prepared by Michael McGoodwin.)

For the past two years I have been writing a book entitled, Open Source for the Enterprise that will be published by O’Reilly later this year and I have been thinking quite a bit about when open source has and advantage over commercial software and when the reverse is true.

I have come up with my own version of Yali’s question: When we look back on the landscape of software twenty or thirty years from now, which type of software will cross the ocean to take over the other? Will open source be the conquering force? Or will commercial software contain the growth of open source to small islands? Or will the map of the software landscape look like a complicated mosaic with each type of software owning certain territory?

To Diamond, winning societies are those that developed food production through farming, supported by domesticated animals, which created surplus food and enabled a sedentary lifestyle as opposed to hunting and gathering. This extra food and time allowed technology to develop because a division of labor led to the creation of specialists. Living in larger groups with animals in close proximity led to lots of germs and diseases being passed around, which created a population of people that evolved to have resistance to certain sorts of diseases and carried germs that were lethal to other societies. As the fundamental causes of who developed food production, Diamond points to preconditions for agriculture, such as the number of wild crops that could be adapted to farming, the number of species available for animal domestication, an east west continental axis that allowed for expansion of crops along the same climate conditions.

One of the interesting things about Diamond’s analysis is how he uses linguistics to track the path of how certain cultures have dominated others. In two chapters, How Africa Became Black and How China Became Chinese, Diamond presents two maps (Figure 16.1, and 19.2 for those who have the book) that show how pockets of languages have sustained themselves in certain areas. Diamond suggests that languages in the pockets used to dominate, but were enveloped by the conquering languages and cultures.

Diamond also presents an interesting analysis of why New Guinea resisted being enveloped by invading Austronesians who landed on their shores but didn’t dominate the island as they did on other lands such as Indonesia, which were populated by hunter gatherers. When the Austronesians invaded, the New Guineans already had farmer powered cultures which enabled them to resist domination, unlike other islands that were still in their hunter gatherer state.

So, which will it be? Is open source vs. commercial software a case of a farmer power vs. hunter gatherers in which the farmer powered culture will dominate? Let’s call this Open Source Wins All. Or will the battle be more equal, farmer-power vs. farmer power? (Lengthy Standoff) Or is the assumption that open source has the farmer-power-like advantage just plain wrong? (Open Source Contained)

Open Source Wins All represents the conventional wisdom among open source advocates. If we think of a map like those in Guns, Germs, and Steel, then in the Open Source Wins case it will look like open source has captured most of the territory leaving commercial software in only remote islands. Byron Sebastian, CEO of SourceLabs, is an articulate advocate of this view. He argues that the “era of extortion is over” and that the “enterprise software business model is dead.” Sebastian sees the software industry as ready to fall prey to a Napster sort of disruption. Commercial software companies will no longer be able to sustain a model of charging a large amount for copying and selling CDs. Sebastian points out that software has become cheap to manufacture because of highly skilled, low-cost workers in countries like India, China, and parts of Europe. Customers for enterprise software are tired of being locked in and would prefer to have choices. Open source will offer those choices and companies like SourceLabs will provide the missing support, quality control, and packaging services to make open source ready for enterprise use. Other companies like SpikeSource and Optaros have been founded on this premise.

But to me, this is a version of bubble thinking which ignores much of the reality of the nature of open source and the nature of enterprise software companies. Commercial software is farmer-powered and has many strengths that open source lacks, which will lead to a Lengthy Standoff. Open source has mostly provided general purpose infrastructure of interest to developers, such as the ability to create a solution for businesses not developers and to combine many solutions to solve the specific needs of an industry. This happens through the product management/product marketing process which looks not to scratch an itch of the person creating the software but to solve the needs of someone else, a potential customer. Commercial companies also are able to create ecosystems of expertise, documentation, and other forms of productization that open source projects do not offer most of the time. I argue in my book that IT departments must over come what I call the productization gap in order to successfully use open source.

Sometimes, some sort of product marketing/product management has brought many different pieces of infrastructure together to create a larger solution. Geronimo is a great example of this. But most of the time open source provides infrastructure that a developer wanted to build. Where is the viable replacement for MS Exchange? Commercial software provides solutions to businesses that solve problems that are of little interest to most developers. Open source is contained by the range of passion that developers have for creating a solution. This range of passion is not as universal as some open source advocates would like to think. Developers are not excited about creating advanced replenishment algorithms for the consumer products industry. They are not going to get peer recognition about writing interface drivers for RFID controllers and then linking them to all of the different warehouse management systems in use.

What about open source companies like Compiere, which is focused on ERP and SugarCRM, which has created products that are well outside the range of passion? These companies are actually commercial software companies that use open source as a marketing tool. They are not open source communities in the same way that Apache is.

So, what will the map look like in the Lengthy Standoff scenario? My view is that large areas of infrastructure will be conquered by open source, but there will be huge area of applications that the farmer powered commercial companies will maintain. Neither side will conquer and many pitched battles will be fought. Byron Sebastian’s arguments about development becoming cheaper apply to both commercial software and open source. Perhaps they mean that commercial software will be subject to more competition, not that commercial software will cease to exist. The commercial software business can learn lessons that open source teaches about software development without having to change the fundamental business model.

In the Open Source Contained scenario we must answer questions such as? What are the advantages of open source vs. commercial software? What is the range of passion? What are the product management/product marketing limits of the open source community process? and other related questions. I leave these to a later blog.

If you want to see a rip-roaring debate of this issue, please come to this session of the Open Source Business Review in which Byron Sebastian, Phil Moore, and a representative of the commercial software industry will debate the Open Source Wins All scenario.

Guns, Germs, and Steel inspired software articles:

  • Geeks, Germs, and Software
  • Guns, Germs, Steel and Open Source
  • Guns, Germs, and Steel Summary

I would love to hear from both the "Open Source Wins All" camp and from those in commercial software companies who have thought about this issue.


2005-06-20 13:35:26
Apache's not all driven by passion
This message comes from a developer who wishes to remain anonymous:

I wouldn't say Apache is 100% passion, there's quite a bit of cash behind the "volunteerism". From my perspective, Apache is a heterogenous mix of motives: maybe 50% of the committers are fueled by atruism, 25% by some giant corporation that has some interest (IBM, Sun, etc.), and the other 25% are eiher independent entrepreneurs
orconsultants. People tend to enter the community from one to the first two perspectives, and then they wise up to the fact that there are a few people in the ASF who use it to fuel highly commercial endeavours usually associated with proprietary software.

For example, Gluecode just made a not-inconsequential amount of money off of Geronimo when that company was purchased by IBM, and many of the people associated with Gluecode where/are core developers of Geronimo and other projects at Apache. This is an example of Open Source
self-regulating this Open vs. Commercial balance.

The Foundation seems to be somewhat agnostic about motive, but they do attempt to guarantee that a project will not become beholden to a specific corporation. Some people dislike this model; it isn't pleasant to contribute to a
project only to see that particular piece of code repackaged and sold for a few hundred thousand. Take my own experience with a Content Management System which contains code that I contributed to the ASF. While I'm satisfied that this company uses the projects I contribute to, I'm also
not excited about working on a piece of software that I could have personally developed

The difference between myself and someone drinking the GPL juice is that I'll go start a company and make a competing proprietary product; I call this "make money selling your own dog food".

2005-06-21 06:29:53
"Lengthy Standoff" with significant inroads
I think that the infrastructure space is clearly being won by open source. Just look at the number of commercial network appliances or security applications that are merely a conglomeration of hardware (sometimes), an administrative UI and a support contract on top of an open source software stack containing BSD, Linux, Apache, or other open source software. Also, look at the low-cost consumer networking devices driven by one of these open source operating systems.

I think that the case is much less clear for enterprise applications. You mention both Compiere and SugarCRM, but fail to differentiate their licenses. Compiere uses the Mozilla Public License without modification, allowing anyone to modify and use their code. SugarCRM continues to tightent their variant of the MPL, making it harder and harder for other companies to independently extend and profit from their code.

Also, your article doesn't mention the growing trend toward true community-supported open source in enterprice applications. CRM applications like XRMS CRM (I am the project coordinator for XRMS) continue to grow with a community of contributing companies that need the functionality. I'm not convined that this will result in hugely popular applications, but it certainly nips at the heals of the large commercial vendors such as SAP, Oracle, Microfost, and Siebel by providing enterprise-class performance at the shared cost of maintaining and extending the software (costs that exist with the commercial enterprise software vendors as well). Moving a high upfront and continuing cost to a low fixed cost, shared marginal cost model may eventually win out in a market that was once dominated by multi-million dollar software sales (with multi-million dollar integration conracts).

Thought Provoking article, Thank You.


- Brian Peterson

2005-06-22 08:11:22
Dan Seltzer and The Wealth and Poverty of Nations
Dan Seltzer in his thoughtful blog ( offers a different book to explain the tension between open source and commercial software.

2005-07-26 06:47:30
Guns, Germs and Steel
I thought you might be interested in seeing my review of the PBS program in July 8 Science. You can access a text version on my Web site at:|Review_of_Guns_Germs_and_Steel.php

all best, Michael Balter, Science