synthesis

A bit of graffiti re Open Source

I dislike the frothing-at-the-mouth ideology (to me, ideology should be something personal, not something you push on other people) and I think it's much more interesting to see how Open Source actually generates a better process for doing complex technology, than push the "freedom" angle and push an ideology.

- Linus Torvalds, in an interview with APC Magazine.

Thoughts on Open Biology

A story at LinuxDevices last year on a report from the Committee for Economic Development (CED), recommending government use of "open source" and "open research", prompted me to collect the following thoughts on Open Biology.

I've changed the entry in my category list for this blog from "Open Source Biology" to "Open Biology".  Despite unleashing the phrase "Open Source Biology" on the world six years ago, at this point I no longer know what Open Source Biology might be.  Perhaps Drew Endy still has a  useful definition in mind, but as I try to understand how to maintain progress, improve safety, and keep the door open for economic growth, I think the analogy between software and biology just doesn't go far enough.  Biology isn't software, and DNA isn't code.  As I study the historical development of railroads, electricity, aviation, computer hardware, computer software, and of the availability of computation itself (distributed, to desktop, and back to distributed; or ARPANet to Microsoft Office to Google Apps), I am still trying to sort out what lessons can be applied to biological technologies.  I have only limited conclusions about how any such lessons will help us plan for the future of biology.

When I first heard Drew Endy utter the phrase "Open Source Biology", it was within the broader context of living in Berkeley, trying to understand the future of biology as technology, and working in an environment (the then embryonic Molecular Sciences Institute) that encouraged thinking anything was possible.  It was also within the context of Microsoft’s domination of the OS market, the general technology boom in the San Francisco Bay area, the skyrocketing cost of drug development coupled to a stagnation of investment return on those dollars, and the obvious gap in our capabilities in designing and building biological systems.  OSB seemed the right strategy to get to where I thought we ought to be in the future, which is to create the ability to tinker effectively,  perhaps someday even to engineer biology, and to employ biology as technology for solving some of the many problems humans face, and that humans have created.

As in 2000, I remain today most interested in maintaining, and enhancing, the ability to innovate.  In particular, I feel that safe and secure innovation is likely to be best achieved through distributed research and through distributed biological manufacturing.  By "Open Biology" I mean access to the tools and skills necessary to participate in that innovation and distributed economy.

"Open source biology" and "open source biotechnology" are catchy phrases, but they have little if any content for the moment.  As various non-profits get up and running (e.g., CAMBIA and the BioBrick Foundation), some of the vagaries will be defined, and at least we will have some structure to talk about and test in the real world.  When there is a real license a la the GPL, or the Lesser License, and when it is finally tested in court we will have some sense of how this will all work out.

I am by no means saying work should stop on OSB, or on figuring out the licenses, just that I don't understand how it fits into helping innovation at the moment.  A great deal of the innovation we need to see will not come from academia or existing corporations, but from people noodling around in their garages or in start-ups yet to be founded.  These are the customers for Biobricks, these are the people who want the ability to build biological systems without needing an NIH grant.

But Drew Endy (Biobricks) and Richard Jefferson (CAMBIA) have as primary customers not corporations, hobbyists, or tinkerers, but large foundations and governments.  The marketplace in which Biobricks and CAMBIA compete for funding values innovation and the promise of changing the world.  At present, they do not derive the majority of their funding from actually selling parts or licenses on the open market, and thus do not rely on sales to fund their work.  Nor should they.  But the rest of our economy operates on exchanges of money for goods and services.  Synthetic Biology will get there some day, too, but the transition is still a bit murky for me.  The Bio-era research report, "Genome Synthesis and Design Futures: Implications for the U.S. Economy", of which I am a co-author, points to the utility of Synthetic Biology and Biobricks in producing biofuels, vaccines, and new materials.  However, the implementation of the new technological framework of genome design, enabled by large scale gene synthesis and composable parts with defined properties, is still in the offing.

Janet Hope has made an initial study of the state of Open Source Biotechnology in her Ph.D. dissertation at Australia National University.  Janet gives the following definition for her project:

"Open Source Biotechnology" refers to the possibility of extending the principles of commerce-friendly, commons-based peer production exemplified by Open Source software development to the development of research tools in biomedical and agricultural biotechnology.

This project examines the feasibility of Open Source Biotechnology in the current industry environment. In particular, it explores:       

1. Whether it would be possible to run a viable biotechnology business on Open Source principles, and

2. What such a business might look like, including the application of specific Open Source-style licences to particular classes of biotechnology research tools.

Janet's book on the subject is due out later this year from Harvard Press.  My book on all of this stuff is, um, not finished.

The CED report  "concludes that openness should be promoted as a matter of public policy, in order to foster innovation and economic growth in the U.S. and world economies."  I think this bit, in particular, is very interesting (quoting from the LinuxDevices story):

• Open Innovation (such as 'peer production' systems like WikiPedia and eBay user ratings)
  
  
  • To foster open innovation, federally funded, non-classified research should be widely disseminated, following the example of the NIH (National Institute of Health)
  • "Any legislation or regulation regarding intellectual property rights [should be] weighed with a presumption against the granting of new rights ... because of the benefits to society of further innovation through greater access to technology."
  • The NSF (National Science Foundation) should fund research into "alternative compensation methods, similar to those created to facilitate the growth of radio, to reward creators of digital information products"

The first point is a bit off, since most NIH sponsored research, as a practical matter, available only through subscriptions to the journals in which it is published.  This will slowly get fixed, however, with increasing publication via the Public Library of Science and similar efforts.  The second point, embodied in patent reform, will probably take forever and will be hobbled by vested interests.  The third may not produce useful results for many years.

So here we sit, needing much fast innovation in biological technologies in order to produce carbon neutral fuels, improve human health, and deal with emerging threats such as SARS and pandemic influenza.  Open Biology is part of that, somehow, but I still don't see a clear path to implementing the ideas within the context of the real economic system we live in every day.

The Costs of Complying with Open Source

The International Herald Tribune has a story by Kevin O'Brien on the costs associated with open source software, "In open source, an unexpected trap".  The future of Open Source Biology will include similar costs.

The article relates several episodes in which companies have included open source code in products without then publishing the resulting code appropriately as dictated by the relevant license.  These infractions have resulted in efforts by coders to push for compliance, and also spawned a new market segment for services that screen for open source code in commercial products.  Palamida, for example, provides code due diligence for a fee.

This is another example of the interesting legal and practical landscapes created by open innovation.  The main message of the O'Brien article for me is that open source continues to be a way for companies to reduce development costs.  And this requires figuring out ways to use open source code effectively, intelligently, and legally.  If code created by the masses is close enough to a solution required by Cisco, Intel, or IBM, it seems the Fortune 500 has no difficulty justifying the use of technology that results from open innovation.  Open source doesn't seem to be killing off traditional companies, as claimed by some large organizations; instead, it's helping the companies that adapt to thrive.  The use of the open source code to reduce costs, and the existence of Palamida, suggest the market is providing the solutions to make open source work.

And if the strategy works for electrons, why not for molecules?  If it works for hardware, why not wetware?  Most relevant to the IHT article, I wonder about verifying compliance with biological versions of open source licenses.  There will obviously be companies spun up to analyze the contents of molecular systems -- genomes, proteomes, in vitro enzymatic cocktails -- just as compliance has become an issue for software companies.

This gets one thinking a bit deeper about the challenges of ensuring compliance.  I suspect open source wetware is like open source hardware, in that compliance probably requires a suite of physical tools that enable one to pick apart the molecular contents of a system unambiguously.  I wrote a few days ago about Intel's Sun's release of the Verilog code for the UltraSPARC T1 chip under and open source license; how are they going to police all the chips out there to make sure some of their code isn't used by a competitor?  Or even in a chip that is used for something else entirely?  If the code for the offending chip isn't published, you would have to subject the chip to all sorts of tests, from running test vectors on the chip to sticking the thing under an electron microscope to directly examine the architecture.

Similarly, looking under the hood of a synthetic biological system to check for open source license compliance will require identifying physical objects and proving their use either is consistent or conflicts with the terms of the license.  Another motivation for better biological test and measurement gear.

If Palamida exists primarily because big corporations don't want to get sued, then I wonder if a biological version -- a service company, say -- can assemble the appropriate tools based on funding from big corporations that want to ensure they are complying with Open Source Biology licenses.  Plus user fees from inventors and developers trying to ensure they get paid?  Interesting.

Sun Tries Open Hardware Development, and India Pushes Open Biology

India is compiling an open, on-line encyclopedia of traditional medical knowledge.  In "India hits back in 'bio-piracy' battle", Soutik Biswas reports for the BBC that, in the last decade, India has found itself working to overturn Western patents on uses of compounds that have been known for centuries by domestic healers.  This prior art is the accumulation of generations of effort, and it is understandable that a population that makes use of traditional medicines might be a tad peeved that their work is being stolen. 

Biswas describes an effort to make the knowledge easily accessible:

The ambitious $2m project, christened Traditional Knowledge Digital Library, will roll out an encyclopaedia of the country's traditional medicine in five languages - English, French, German, Japanese and Spanish - in an effort to stop people from claiming them as their own and patenting them.

A major motivation for putting all this information in writing is that the oral component of traditional Indian teaching and knowledge is not acknowledged withing Western Intellectual Property law:

Under normal circumstances, a patent application should always be rejected if there is prior existing knowledge about the product. ...But in most of the developed nations like United States, "prior existing knowledge" is only recognised if it is published in a journal or is available on a database - not if it has been passed down through generations of oral and folk traditions.

There is obviously a great deal of value in this accumulated wisdom:

Dr Vinod Kumar Gupta, who is leading the traditional wealth encyclopaedia project and heads India's National Institute of Science Communication and Information Resources (Niscair), reckons that of the nearly 5,000 patents given out by the US Patent Office on various medical plants by the year 2000, some 80% were plants of Indian origin.

By one estimate, a quarter of the new drugs produced in the US are plant-based, giving the sometimes much-criticised practitioners of alternative traditional medicine something to cheer about.

Which suggests an additional effect of the library: those inclined to self-medicate will have a tremendous resource for treating what ails them.  We are likely to see an increasing number of people showing up at western clinics and hospitals with a history that includes treatments and compounds not amongst the recognized armamentarium.  Another complication of Open Biology, or Open Source Biology, or whatever we are going to call it.  People are going to use information however they see fit, trying things out, producing improvements occasionally.  It's another matter, of course, as to whether those improvements will be shared.  One can only hope that the tradition of open innovation extends to such novel medical treatments.

Sun Microsystems appears to be explicitly counting on this behavior to provide improvements in their UltraSPARC T1 processor.  By open-sourcing the VERILOG design code (eWeek news story) for the chip (OpenSPARC), Sun is hoping the masses can produce more innovation than Sun itself.  The press release makes interesting, if flowery, reading.  "If it works in software, why wouldn't it work for processors?" asks Chairman Scott McNealy, as quoted in the eWeek article.

The most interesting part of the whole story is the strategy to promote innovation around the chip, and then potentially bring those innovations in house.  Jeffrey Burt, in the eWeek article writes:

Sun already has shown the ability to bring in key technologies through acquisitions—indeed, the groundwork for the T1 chip was developed by another company, Afara Websystems Inc., which Sun bought in 2002. McNealy said he envisioned a future where companies will be created to develop technologies around UltraSPARC T1, and then be acquired by Sun. (emphasis added)

If this strategy works for hardware, why not wetware?  If it works for electrons, why not molecules?  I speculated about this in an earlier post, Acquiring Open-Source Projects.  It appears that with the OpenSPARC project we will have another example of how to encourage innovation, and we will find out whether a commercial entity can profit from so explicitly sharing the fruits of its labor.

Acquiring Open Source Projects

One of the issues that always arises in discussions of Open Source Biology is how anyone will make any money.  Not everyone is interested in money, of course, but molecular biology is still a bit expensive and requires capital of some sort to keep it running.  Without the possibility of a return on their investment, most investors probably won't go near open source biology.

This was an explicit objection raised by a banker/VC type at After the Genome VI, in December of 2000.  I've heard similar complaints all along the way, though the VC in question also added something like "they won't let you do it" to his oration, presumably refering to big biotech companies.  I immediately asked myself what "they" could do about it, and have pondered the question ever since.  The problem doesn't seem to be encouraging tinkerers to have at it, or that big companies might prevent them from doing so, but rather turning the fruits of tinkering into useful tools and products that most people want to use.

Products take a long time to develop to the stage where people want to actually purchase them.  Lot's of open source software in particular doesn't attract a large user base because the interface isn't as polished as that provided by commercial houses, even if the guts of the code are better.  This is as true of software and cars as it is of molecules and other biological technology.  In my experience, most biologists seem to want a box with an instruction book -- a package they can use to produce data -- and are rather less likely to put up with sorting out the intricasies of working with a tube full of molecules from some guy down the street.

On the one hand tools and skills are proliferating at a remarkable rate, democratising the technology and its applications, but on the other most new useful tools still come from "traditionally" funded and run corporations, and we want to ensure continued investment that funds that development of finished products.

One way out of this might be the aquisition of open source projects by established companies, or by start-ups funded specifically to take a project private and push the commercial applications.  It turns out this has now happened in the open source software world.  This obviously can only work if all the contributors to the open source project agree to sell their rights as developers to the company.

David Berlind describes what transpired, and explores its implications, at ZDnet:

To acquire an open source project, the acquirer must be absolutely certain that they are acquiring the copyrights to all of the code being used in the project.  Those copyrights ultimately belong to the individual contributors to the project who, up until the point of acquisition, would have been bequeathing certain rights to their code to others under whatever open source license is behind the project.  To the extent that licensing that code under an OSI-approved license is what let the code out out of the box and into the open source wild, there’s nothing that the acquirer can do to put it back in the box.  That code will always remain available under whatever open source license it was published.  But, by acquiring the copyrights and any trademarks associated with that code, the acquirer also acquires the right to modify and distribute the original code without having to make those modifications available under an open source license.  In other words, future versions of the open source software could become closed source.

The last sentence is perhaps the most interesting, particularly in the context of biology.  I can imagine open source biological technologies developed in a distributed way, or at least developed by more than one person, which are useful to those willing to master the eccentricities but which are not widely used because they may be unwieldly.  In steps a commercial endeavor to tie up all the loose ends, and then put it in a nice package with a bow on top -- complete with instruction manual, please.  As with software projects, all the details disclosed prior to aquisition would remain in the public domain, but any further work the company put into development would remain their property and contribute to the value of the final product.

This is, of course, similar to how technology is moved from universities into the private sector.  So it isn't a great stretch of the imagination to see that it might work with distributed, "amateur" development efforts.  Something to consider.

East Bay Express Article on Open Source Biology

Here is Nathanael Johnson's East Bay Express article on the origins of Open Source Biology.  I don't know if we deserve as much credit as Mr. Johnson gives us.  But I do miss hanging out at the Palace...

UPDATE (18 April 05):  Here are the original Intentional Biology and Open Source Biology web pages, which haven't been edited since 2000 or so.  Content by myself and Drew Endy.  This material is basically ancient history at this point, but not so bad for following the development of the ideas.