January 04, 2005
Does the open source model apply beyond software?

Does the open source model apply beyond software? This is a question being asked more and more. For instance, is there potential for "open source biology"? If I actually knew biology, I might be able to answer that question. However, in twisting my brain to say anything meaningful, I had a Eureka moment.

I'll go out on a limb. Open source heralds a global paradigm shift in social and economic value creation of enormous proportions, the extent of which is almost completely unappreciated. If I am right, then we are in for interesting times as the irresistible force of open source meets the immoveable object of corporate entrenchment. There will be enormous economic dislocations with opportunities for new undertakings and threats to the established order. Mental categories we take so much for granted we don't even know we have them will become obsolete.

Part of me doesn't want to say this, even though I believe it, because too many times I've regretted my public stands saying that revolutionary change via technology (the PC, the Internet) was at hand. Probably there is something deep in my personality, reinforced by growing up in the 60's, which gives me a predilection towards issuing manifestos. My adult self thinks it is indeed wise to speak with some restraint. So think of this as bold speculation designed to provoke serious thought and reaction, not as prediction.

A final aside - I care about these matters because I like and appreciate the good side of business so much. After all, I have been an enormous beneficiary of the system, and there is something great about the idea of starting from nothing and crafting a business based on inspiration and hard work. (Yes, I know I am simplifying and mythologizing here. It's one thing to start from "nothing" as a well-educated child of a middle class family in the U.S. It's another to really start from nothing as a child of poverty in Africa or anywhere in the less developed world). At the same time, the standard practice of business is pretty disgusting when, as it often is, it's about greed and hurting the competition more than creating value and succeeding on one's own merits.

Business, writ large, is the archetypal form around which our world is organized. Its chief pillars are private ownership of assets and market-based competition. Despite the prevalence of self-serving mischaracterizations and misguided norms, e.g. that free markets are really free, and that the best government is the one that interferes the least, "the business system" is responsible for the prosperity and opportunities we enjoy, as well as problems and miseries it creates.

To say, then, that open source represents a fundamentally new and different way to organize economic activity which challenges the dominant paradigm, is not a modest claim.

Many of the best efforts to characterize what is fundamental about open source partially miss the mark. It is decentralized, a point a June 2004 Economist article (buried in their non-free archive sadly) captures well. To say that the other central characteristic of open source is that source code is freely shared, while true, misses a central point.

Open source is a form of decentralized production in which an information commons is cooperatively built, maintained, and evolved. This commons forms the primary (but not exclusive) basis of value from which individuals and firms draw and to which they contribute.

The centrality of the commons in an open source paradigm contrasts with private ownership of assets in the existing one, and the dynamic of use and fostering of the commons stands in sharp contrast to the dynamic of market-based competition.

When information such as genomic data is put in the public domain, it may be freely used, but that by itself does nothing to foster a dynamic of further contribution, e.g., more data, refinements of the data, tools for analyzing and manipulating the data, etc. Open source projects have dynamics of active contribution and community participation over time as a distinguishing characteristic. Sometimes these are enforced by the software licenses employed by the project, the Free Software of GPL license which compels users who create derivatives to contribute them in their entirety to the commons, but this is not necessary as many projects do not rely on compulsion to support community contribution. I am ignorant of current practice, but I would look at information sharing going on in biomedical research as to whether there is an active community-based dynamic going on or not, and if not, whether there could be.

Empirically, I would say the fostering of a common results in improvements in products over time, which we ordinarily assume requires competition to achieve. An important unsettled question about open source is whether larger innovation, as opposed to incremental improvement, can also happen this way. This is a point the article makes.

However, we tend to romanticize and sometimes overestimate the centrality of big breakthroughs in the scheme of things, short-weighting the less glamorous processes of incremental innovations which actually count for a great deal in my view. Another way of framing is this issue is to ask to what extent major innovations need to be motivated by the prospect of creating vast wealth through private ownership and control.

More to comeā€¦

Posted by mitch@osafoundation.org at January 04, 2005 08:46 AM

Or it can be argued that Open Source is a return to the traditional methods of science and exploration. Certainly the old groups like SHARE, CACM, Usenix, etc. had a strong sense of academic sharing of techniques and results. If you go to much older domains like the European exploration of the Americas you find a similar mix of open disclosure and secrecy regarding the new lands being explored.

I view Open Source as just the continued evolution of old scientific attitudes under the influence of new technological capabilities.

Posted by: rjh at January 4, 2005 11:25 AM

I agree with rjh that open source works a lot like science --- and this is probably because a common aim is the world-wide accumulation of abstract things (knowledge or useful software).

I don't know just how familiar you are with academia, but if you are a member of a wealthy research institution (usually a university), then you have free access to almost all available published research. This makes it possible to find and learn what other researchers have discovered. Independent researchers or those belonging to poorer institutions have a much more difficult time of things (though the internet is improving that a bit). Regardless of your affiliation, any researcher can freely make use of other researchers' results (once they know of them), and can also mimic and then modify other researchers' experiments or reasoning --- so there is a clear analogy to open source.


/Free access/ and /unrestricted rights/ are vital to the /accumulation/ of knowledge (and useful software). Unfortunately, I believe that those 3 things (free access, unrestricted rights and, importantly, accumulation) conflict with the aims of businesses based on selling IP. More disappointing is that many capable businesses (that don't sell IP) don't contribute to public research or open source software, because their efforts become available to competitors. It makes sense: if an innovation is available to all businesses it means no business can improve their revenue.* But the innovating business will have spent resources to develop the innovation.

As for innovations, I've always believed the biggest innovations come from science, not industry. This is because scientists pursue science for, and are normally funded (usually by governments) for, the sole purpose of discovering new things. But businesses will also innovate if they believe there's an unmet public desire they can profit from. If you compare these situations with open source, you can see why open source struggles to produce major software innovations (where the code is often not an end but simply a way to achieve a goal, and the 'scratch your /own/ itch' philosophy reigns). Costly innovation in open source is possible, but not likely (i.e. not frequent). On the bright side, dedicated and non-profit organisations can buck this trend.

I think open source can be very successful when: 1) development and distribution costs are very low (relative to most industries); 2) benefits can scale to any number of users without any increases in cost; and 3) those benefits accumulate over time. For example, Wikipedia can work well because it breaks up a huge and costly task into lots of smaller and less costly tasks. OTOH, most movies and books today break rule 1 and usually 3 (but if you can invent a distributed or cheap way of writing or generating a movie or book . . .)

There is so much that should be discussed about these issues --- it's a pity politicians rarely consider them.

* Assuming the businesses can't use the innovation to attract wholly new customers.

Posted by: voracity [TypeKey Profile Page] at January 4, 2005 09:36 PM

how about "open source *architecture*"...? http://www.haque.co.uk/opensourcearchitecture.php

Posted by: charles at January 5, 2005 04:26 PM

Biology presents a number of intriguing generalizations of the open source idea, for instance:

>>BIOS (http://www.bios.net) is literally attempting to replicate the SourceForge model for bio research, especially in areas where markets are failing to produce results.

>>BioMed Central (http://www.biomedcentral.com) is part of an "open access" journal movement to open source research results, especially to poor countries that can't afford the price-gouging of the big publishers.

>>GenBank (http://www.ncbi.nlm.nih.gov/Genbank) is a classic success story of an info commons. It has partly been fostered by the requirement from journal publishers that researchers deposit sequences in GenBank prior to publication.

However, this doesn't address your basic question, which is whether "open source represents a fundamentally new and different way to organize economic activity"? I don't know the answer, but would propose approaching it from the opposite direction: Does the open source model have something essential to teach us as we try to move towards a sustainable economy? I think it may, given the fact of the "centrality of the commons" to it and the fact that the looming ecological crises that we face are in large part failures of our present economic model to account for the commons.

Posted by: sharp11 [TypeKey Profile Page] at January 6, 2005 07:39 AM

Open source is also similar to the very traditional business model of "teaming" for consortia, limited partnerships, etc. Although it is based in the federal copyright law, the GPL can also be viewed as a form of teaming agreement. Unlike the older teaming models, there is a tiny transactional cost barrier to joining an open source team. The older models usually involve joining some sort of organization, with fees, rules, etc. Joining an open source "team" can be done unilaterally, provided you agree with the team rules. The CACM and SHARE were similar earlier efforts, with relatively low cost of entry.

Teams have long been a good method for infrastructure development. It is generally easy to get the participation of the infrastructure users in a team to create infrastructure, provided the cost of entry is low. The participants view participation as a means of lowering their overhead costs. Infrastructure is not a proprietary interest to its users. The primary opponents of such teams are the infrastructure vendors, who view infrastructure as a profit source rather than an overhead cost.

This is part of why open source has done so well in infrastructural domains like operating systems, networking facilities, email, etc. As specialized applications become generic infrastructure, e.g. word processing, open source should do well there.

For these kinds of businesses open source is a substantial reduction in transactional costs that does have a substantial impact on infrastructure and what is considered infrastructure. It is not a fundamental change in the business model. I would compare it to the changes in banking caused by automation, ATMs, etc. Banking fundamentals are still banking fundamentals, but the practice of banking has changed significantly.

Posted by: rjh at January 6, 2005 08:48 AM

I would reinforce the point about traditional scientific practices from my experience with agricultural research. Up to the late 1980s, the tradition was the free exchange of both research results and actual germplasm (seeds) that resulted from crossing plants and other genetic manipulation. Great seed banks were run with public domain germplasm at government institutes and universities world wide. Some important agricultural varieties were privately developed and patented, but they were far outnumbered by those in the public domain.

Then intellectual property protection was expanded just as genetic engineering provided new tools for invention. The result has been an very complex patchwork of private rights and limited distribution of germplasm and research results. Many biologists I know in international agriculture want to create more open endeavours, in a kind of backlash to the new order.

Posted by: ken novak at January 7, 2005 12:44 AM

Dear Mitch,

You wrote: "Open source heralds a global paradigm shift... There will be enormous economic dislocations with opportunities for new undertakings and threats to the established order."

"Will be?" I can't speak to economic dislocations, but in regard to "threats to the established order," this is no "will-be" pipe dream. It already a cornerstone of daily life.

For instance, see:

"Amateur Revolution"

Speaking from personal experience, I know that one can capitalize on two extremely powerful forces to advance the state of the art in a field of study: the generousity of countless writers and thinkers (many of them unpaid and underappreciated) who post their findings freely to the Web for easy access by all comers; and Google, which effectively acts like a virtual research staff of hundreds of expert librarians, locating these resources for me.

By leveraging these, I have found it possible to answer questions about the origins of "Leaves of Grass" that have never been adequately addressed by the cartel of scholars who effectively defined and limited the range of permissible topics, opinions, and approaches.

Yes: everywhere, and constantly, we are indeed threatening the established order.

Mitchell Santine Gould
Curator, LeavesofGrass.Org

Posted by: GeneralPicture [TypeKey Profile Page] at January 8, 2005 07:19 PM

rjh wrote: This is part of why open source has done so well in infrastructural domains like operating systems, networking facilities, email, etc.

rjh, you're close, but go further. To the average company or user outside the software industry, *nearly ALL* software they use is just "infrastructure". E.g. selling taps and shower heads to the public. There are probaly one or two companies doing ground-breaking work with real-time customer portals showing personalized plumbing styles in 3-D with real-time order histories and shipping details, and super-whizbang internal ERP/EAI/CRM systems supporting this. But for every other plumbing company, they would be better off participating (or having their IT contractor participate) in open source solutions for all their software needs and spending their money elsewhere. And I believe that in a few years, the open source solutions in *ALL* those areas will be good enough such that people and companies who focus their expertise elsewhere than software will suffer little competitive disadvantage.

Posted by: skierpage [TypeKey Profile Page] at January 9, 2005 03:20 PM

To my mind, a potent example of "Open Source" with lots of obvious tangible benefits is the "alternative health market". On the Internet, there has to be thousands of health regimes, medicines, treatments that are worked on collaboratively, iteratively and with complete open intellectual license. Moreover, it has the same ramifications for those that dare to build their own solutions. Namely, there is ample evidence that the orthodox health industry wishes to stamp it out, such as the threat to close down the supplements industry in Europe, apparently because such things might be "dangerous" for our health.

It seems to me that this industry can be made even more effective with Internet tools that are designed to allow medicinal collaboration to become more effective. In other words, applying open source tools more formally will reap benefits.

For example, in many cases it seems that careful logging of treatment regimes and symptom observations is essential, but seldom done in a systematic fashion. Even mobile technology could be put to good use here to allow easier and more real-time logging to occur. Mobile technology can also be used for better communications between "patients" where mutual support networks are a vital part of the process. Moreover, many treatments have subtle variations that patients come up with that work in their case. Effective gathering of this information in a systematic fashion would assist other collaborators ("patients") to "download" the most appropriate solution for their situation and then to feedback their experiences and variations into the appropriate "development thread". There are direct parallels here with source code control.

In effect, the Internet is allowing new "virtual doctors" to emerge. Many of us will have our own GP (physician) but perhaps several other virtual ones too. The treatments themselves are also available - without costly IP license costs. I am confident of a revolution already underway, possibly of greater and more profound impact than the OS paradigm shift in the software industry.

Posted by: Paul Golding [TypeKey Profile Page] at January 11, 2005 12:56 PM

The body of the Open Source revolution may be software, but the blood of the Open Source revolution is trust. That's the same blood, when you think about it, that flows through the veins of healthy economies, healthy communities, healthy families, etc. When there is a foundation that allows trust to occur and to florish, great things happen.

Wikipedia is about trust. We trust that people who add to it are sincere and knowledgable. We also trust that others are motivated to monitor Wikipedia and weed out the wankers who aren't mature enough to deserve our trust. I tried to create a Wikipedia entry last fall, based on a neologism that sprang up during the presidential campaign. I thought that "Certsian Philosophy" was a cool turn of phrase that needed to be documented when I saw it in use, but the guardians of Wikipedia deemed it not worthy by WP standards. At first I was sore, but then I took the long view and realized they were doing the right thing for the common good. I know I can trust these unseen people to continue to do good by Wikipedia, because there is a framework that both attracts trust and allows that trust to prosper.

I just downloaded Chandler because I trust that anything with your name attached to it is worth taking a look at. (I hope I'm right).

I'm not a Trekkie, but I know that when some villian does evil in Star Trek, the characters often act as if they have been personally violated, because the Federation culture is essentially built on trust. (SF's portrayal of an open source culture?)

Trust is not an act of blind hope; it is a covenant. When we trust each other (one-on-one or in a community), we are acting out on a committment of shared values. I use Foxfire and Thunderbird because I know I can trust the community that created them. I post bug notes (rarely -- I don't use real early stuff) because I recognize that trust is a two-way street.

Taking the Open Source revolution into other endeavors means finding a mechanism to put trust into the relationships of that particular endeavor.

Posted by: Randall Newton [TypeKey Profile Page] at January 13, 2005 08:21 PM

Here's an interesting perspective. I think it was 1998 when Eric Raymond started publishing the Cathedral and the Bazaar (book). He briefly pondered the question of whether the open source approach would work outside of technology, but turned away from the question saying that we needed to prove that it works in the open source arena (I'm paraphrasing from memory here)... but that we should consider such question in 5 years i.e. 2003.

So here it is 2005 and Mitch is going out on a limb, to pose the question... which if true proves the part where Mitch said "the extent of which is almost completely unappreciated. If I am right, then we are in for interesting times as the irresistible force of open source meets the immoveable object of corporate entrenchment."

You are right, Mitch. Ah to live in interesting times, eh?


Posted by: eno [TypeKey Profile Page] at January 16, 2005 05:35 PM

This typekey thing decided I'm spanish and does not allow me to chose any other language, so I cannot log in (what does "usuario", "correo" and other prompts mean?). I had to hack the page to post this.

The reason why the open source model cannot succeed in areas where it is not already a success is because for sw development and wikipedia you only need "nothing" (a PC, an internet connection and some specific knowledge that you earn your living with or is your hobby) to create something.

For, as in the examples cited, biology, you need money for instruments, guinea pigs an reagents. So if those are part of your "nothing" because you work in a publicly founded research center, you can freely contribute (i.e. hijack and peruse contributor's money). At least until your boss is going to ask you the list of scientific papers you've published, so he can ask for next year's funds.

Unluckily the laws of thermodynamics always work. Open source is not and cannot be a perpetual motion machine...



Posted by: Guido Gambardella at January 27, 2005 02:26 AM

Transparency, not trust, is the driver of the open-source perpetual innovation machine. Open-source is distinguished from communism by precise definitions of process transparency. This transparency reduces political and thus economic risk. Not every industry can replicate open-source process mechanics, but all industries can achieve the risk reduction benefits of process transparency and social learning mobility.

In the reposted text below, FSB = Free Software Business, FSF = Free Software Foundation. [ Source: http://www.crynwr.com/cgi-bin/ezmlm-cgi?mss:8056:200209:hlpdhbfgkkkhjgnapmgj ]

-- reposted text --
A traditional rationale for proprietary software pricing is the need to fund innovation, i.e. R&D. From that perspective, customer selection of a vendor equates to a vote of confidence in the vendor's R&D strategy. Secondarily, there is recognition of the customer's increased voice during triage of future requirements.

What is the non-zero-sum value proxy that is traded in the FSB ecosystem, and what is the perceived return? It's not innovation, which occurs with religious spontaneity. It may, boringly enough, be survival. A customer decision to invest (funds, attention, packaging, performing) in a niche of the FSB ecosystem is a vote for risk reduction in that niche.

Independent of the ethical and moral subtleties of relative freedom, there is one benefit that accrues to all participants in any ecosystem. In an FSB ecosystem, that benefit is particularly valuable, because of the extreme pace of innovation in FSB ecosystems.

That benefit is the "Existence Proof", otherwise known as the great Taxonomy Debate of librarians and codeline managers. The self-recursion of
taxonomies (incl. software) is broken by the existence of a customer who applies the content/software/media object in the context of a finite geography and finite "offline" market.

It is the application of infinite-innovation FSB products to finite, traditional markets that returns value to the FSB ecosystem. When everyone is mobile, stationary objects have economic value. When all products are evolving at a frantic pace, stable application requirements have economic value. When process transparency accelerates process innovation, economic value shifts to the latency of contracts.

Innovation, the strength of FSBs, increases what might be called marketing costs ... in all stages of the pipeline between R&D and shipped product.
What happens when the half-life of an FSB product becomes shorter than the most efficient execution of the marketing pipeline? You have to cut
"identity costs". Branding (visual promise to execute consistent process) reduces identity costs. So does bundling and consolidation.

Transparency of source code permits FSB distribution channel participants to wander through all syntactic levels (mailing lists, CVS, patchsets, licenses, gossip) of FSB product delivery. But without the cultural and semantic context required at each level, the social mobility afforded by FSBs cannot be exercised by most ecosystem participants.

Nevertheless, like other civil rights, FSB social mobility can be valued by all participants, using a visible subset of individuals as proxies. The regular soap operas in various FSB cores provide reassurance to those in FSB retail that freedom is alive and arguing. Making core debates visible at retail is critical for succession throughout the ecosystem.

The longer the mean distance between FSB cores and FSB retail, the better. From a business perspective, may they grow ever further apart. That creates an entire topology for value mediation, advocacy, arguing and yes,
mobility. As long as the FSF (or other watchdog of boundary conditions) can provide existence proofs for transparency and mobility along _an_
entire FSB value chain, we need not restrict the length or opacity of _all_ FSB value chains.

Posted by: Rich Persaud at January 28, 2005 04:56 PM

We are running a record label at http://www.locarecords.com that tries to use ideas from the open source and copyleft community. We are using the Creative Commons licenses to do this. Feel free to drop by...

Posted by: David at January 29, 2005 05:37 AM