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.