Aaron Ault, James Krogmeier, Dennis Buckmaster 2018-02-16 23:34:59
There are some common laments in agricultural data that seem to remain despite repeated attempts to address them: lack of interoperability, varying data privacy, varying data quality, and difficulty in attracting talent from a population increasingly disconnected from agriculture. At first glance, these issues seem unrelated to each other: how could solving interoperability issues relate to attracting talent? How can issues of data privacy affect data quality? Why is agriculture finding it difficult to attract talent when it offers some of the most interesting, challenging, fun, and relevant problems in the world? The common thread of all these problems is that their solution can be found in open source culture. It is important to address one misconception up front: “open source culture” does not mean “make all private agricultural data publicly available.” Open source refers instead to a method of software development in which the source code is made available publicly without requiring a licensing fee. The data, which the software handles, remains as private as the owner wishes it to be. Only the code is open, not necessarily the data. Of course, once open source code makes people masters of their own data, they are empowered to make their data available as they see fit, be they researchers, industry, and even farmers. The hassle and time sink of wrangling data should not be the limiting factor in getting value from the data. The open source movement began as a novelty in the tech industry and academia. At the time, few people thought that open source offered real potential for profit or progress. The basic idea was that volunteers would share their source code for free in hopes that someone else would find it useful. As competition grew to create ever more useful tools, communities formed around some of the ideas that had the broadest applicability or most engaged proponents. However, beyond a few academics and software engineers with spare time, how could anyone justify the rigors of software development with no clear path to revenue? Who was going to undertake the boring tasks, like writing documentation or performing quality control testing, when it was more fun to code up a buggy descendant of Tetris with candy cane lasers? What happened in subsequent decades has been miraculous. The fledgling community of open source volunteers has grown into a movement that eclipses all others as the basis for nearly every piece of software written anywhere in the world. Technology companies that ignored or fought against open source for years have either disappeared or transformed themselves into ardent open source supporters. How this happened is beyond the scope of this article, but why open source turned out to be a huge success is worth mentioning here due to its relevance to agriculture. Open source democratizes innovation Innovation has been described as an evolutionary process: previous successful innovations mutate and fuse, giving birth to new ideas. Through varied methods of natural selection, some new ideas die while others survive to supply the genes for future generations of ideas. The speed and efficiency of the innovation-evolution process are determined by the number of new ideas simultaneously living, the speed and accuracy of the merit-based elimination of bad ideas, and the extent to which the ideas of tomorrow avoid retrying the failures of the past. When the keys of innovation are held by only a select few, this evolution toward “better” is handicapped because fewer people trying fewer alternatives means that fewer ideas are born. Fewer people in the selection process also means that the most creative ideas are kept down because they disrupt established norms, and a meritocratic evaluation of ideas falls prey to office politics. Open source democratizes this process. Anybody with an idea, passion, and talent anywhere in the world can contribute to the innovation evolution. More minds mean more ideas and more creativity. More evaluators across more cultures mean faster, more effective natural selection for a wider variety of uses. And more widespread geographies force more rigor in communication and process management. As an analogy, we’re all familiar with how inbred genetic lines limited agricultural production in the past. Since then, we’ve learned how to feed the world by harnessing the vigor of hybrids. Open source reduces the barriers to building on others’ work As researchers at all levels can attest, it’s difficult to build on the work of others when all you have is a published paper. You can spend months attempting to recreate a vaguely described algorithm, only to find that the end of that development rainbow is not a perfectly reconstructed pot of gold but rather a fuzzy facsimile that lacks an objective standard of comparison. With open source, researchers do not start with what the author says about his or her work. They start with the actual work—the actual code, the actual test cases, and even the actual data. Day 1 begins with real innovation, not reinventing a wheel that someone else already made beautifully round. Releasing your research as open source makes it more valuable to others, and therefore more highly cited. Open source builds bigger markets Probably the most compelling part of the open source movement is how the tech industry found more profit in combinations of open source and proprietary development than in proprietary development alone. It turns out that the best way to mobilize an industry toward creating bigger markets is through open source. If Linux and Apache had not become the internet’s open source backbone, then a closed, proprietary internet could have taken over—or a suite of mutually incompatible parallel internets: a Microsoft internet, an IBM internet, an Apple internet, etc. Fortunately, owning 100% of a small proprietary internet generates minuscule profits compared to owning just 1% of the actual open source, open collaboration internet. For example, France Telecom’s Minitel system was born a decade before the World Wide Web was introduced. It was open to a degree and successful in its day, peaking at about 20K online services in the early 1990s. However, it proved just a bit too proprietary, given that its backbone architecture was controlled by a monopoly telecommunications company. Ultimately, it was steamrolled by the open source, open collaboration internet. Put another way, you make far more money when you treat open source as a powerful ally rather than as a competitive threat, because we create more value collectively than anyone can create alone. Agriculture is currently experiencing the problem of small proprietary markets. Many companies have dedicated themselves to making their own proprietary internet instead of working together to build an ag internet. Some companies restrict how people are allowed to use their data across platforms. Some argue with farmers about who owns the data. Some treat ideas that originated outside their company, or in the academic world outside their research group, as useless at best or usually problematic. Unless open source collaboration can be brought to bear in industry and academia, we will all continue to lament the limited utility and limited size of the ag data market. Open source streamlines talent discovery and attraction When Facebook released its now wildly popular React.js open source framework, one of its lead engineers mentioned in his keynote address that he “felt like Steve Jobs up here.” You could see the pride and excitement in his face. He and his team had made something really cool, and he got to tell the world about it. Keeping our ideas secret prevents our current talent from inspiring the talent of tomorrow. If agriculture wants to attract the best people, we need to create this spirit, this excitement, this culture of discovery in our companies and our universities. We face the grand challenge of feeding the world’s growing population, now and into the future. It should be easy to inspire talented people to commit themselves to this important work, rather than creating yet another social media app. If we sponsor and participate in open source around the world, then those talented people will appear, and they will want to work with us. Our open source success stories In January 2018, we officially launched the Open Ag Technology and Systems (OATS) Center at Purdue University. Our industry partners, including WinField United, Centricity, ADM, AgGateway, and CNH Industrial, as well as a key grant from the Foundation for Food and Agriculture Research, are providing the funding. Our goal is to partner in a long-overdue open source community in agriculture through projects, education, collaboration, and communication. Over the past few years, our OATS group at Purdue University has been working on data, software, and interoperability issues. We’ve had the privilege of learning from a diverse group of excellent people across broad swaths of the agriculture industry. We recently had an amazing experience at a hackathon run by some great people from the FarmHack community (https://farmhack.nl). Our winning solution was an open source framework and apps for the Dutch swine industry. It uses real-time private data connections coupled with data science to optimize genetic traits based on actual onfarm outcomes. The foundation for this idea arose from our Open Ag Data Alliance project (http://openag.io), which has found new application in the recently released Trellis framework for produce safety data (https://trellisframework.org). We are convinced that open source can bring more rapid solutions like these to agriculture. It’s time to stop lamenting data problems and start solving them. We invite you to join us in a worldwide agricultural open source community to find innovative ways to make data work better for agriculture. Aaron Ault, Senior Research Engineer in Electrical and Computer Engineering (email@example.com), ASABE member James Krogmeier, Professor of Electrical and Computer Engineering (firstname.lastname@example.org), and ASABE member Dennis Buckmaster, P.E., Professor of Agricultural and Biological Engineering (email@example.com), Purdue University, Lafayette, Ind., USA. All three authors are founding members of the OATS Center, an integral part of Purdue’s Digital Agriculture Initiative, which seeks to improve agricultural processes through more efficient use of data.
Published by ASABE. View All Articles.
This page can be found at http://bt.e-ditionsbyfry.com/article/Ag%E2%80%99s+Future+Belongs+To+Open+Source/3011949/476169/article.html.