Intellectual monopoly-based business models [1
], which rely on patents [9
] and their concomitant challenges [8
], are common across both hardware and software industries. However, more recently, the software industry has embraced the concept of liberating otherwise restrictive IP using Free and Open Source Software
]. FOSS is computer software that is available in source code form and can be used, studied, copied, modified, and redistributed either without restriction or with restrictions only to ensure that further recipients have the same rights under which it was obtained: free, or libre. Here the term “libre” will be used rather than free to convey the freedom (of access, use, and discovery) that comes with free and open source materials, not only the lack of cost ($
0). A substantial literature now exists on the benefits of FOSS over historic and more established models [14
] and as shown in Figure 1
the concept is covered at a high rate in academia. The secret of the libre approach is that large-scale collaboration on technical problems results in superior design and innovative solutions with lower associated costs due to continuous improvement [14
]. It is now well established that FOSS is more reliable and relevant to users [23
] (in part because many FOSS users are co-developers [24
]). FOSS has become so prominent in the software industry that it represents a significant change in the career trajectory of software developers [25
] and is dominating major areas of computing. For example, Android, an open source-based operating system, is the world’s most popular operating system [26
], 97% of the world’s supercomputers use GNU/Linux [27
], and major Internet-based corporations use and develop FOSS including: Amazon, Alphabet (Google), Twitter and Facebook. These companies and others use FOSS because of superior technical performance, more flexible design and reduced research and development (R&D) costs [28
Historically libre technical development was limited to software as the FOSS community believed that libre hardware was challenging because of the necessity of building physical objects [29
]; however, within a few years these views changed with the introduction of low-cost digital distributed manufacturing tools [30
] such as the self-replicating rapid protoyper (RepRap) 3-D printer family [31
]. As FOSS has proven so successful, free and open source hardware (FOSH) gained traction in the technical communities [35
]. FOSH is hardware whose design is made publicly available so that anyone can study, modify, manufacture, distribute, and sell the design or pieces of hardware based on that design [36
]. FOSS/FOSH-based technical development provides a competitive advantage as innovation can include more participants than proprietary innovation within firms [14
], it is less encumbered by IP issues [37
], and innovation occurs at steeper rates [40
]. The scientific community has been quick to adopt libre hardware as it reduces costs by 90–99%, while improving control and allowing customizable features [41
]. In addition to traditional publishing, new academic journals such as HardwareX
(publishing libre hardware designs from potentiostat/galvanostats [44
] to a microsyringe autosamplers [45
]) and the Journal of Open Hardware
, which covered the annual Gathering for Open Science Hardware [46
] and libre hardware-based business models [47
]. Some libre hardware designs have been shown to be growing at an exponential rate [41
]. In addition, as shown in Figure 1
, the academic community is rapidly embracing libre for both FOSS and FOSH, with the latter lagging by about 20 years based on indexing in Google Scholar.
Companies embracing libre product development range from the multi-billion-dollar RedHat [48
] to hundreds of small startup companies [35
]. Most (at least 78%) of companies now use some form of open source product [49
]. Companies often partner with universities to do research following different models such as collaborative research, university-industry research centers, contract research, and sponsored research [50
]. Companies generally begin their interaction with universities using a proprietary standard research agreement (PSRA), which for the historical reasons described above contains significant IP monopoly language and restrictions for both the company and the university. Such standard research agreements thus create an artificial barrier to collaboration as both companies using a libre model and universities they wish to collaborate with must invest significantly to restructure the contracts. To solve this problem, this article provides a new Sponsored Libre Research Agreement (SLRA). The differences between the PSRA and SLRA are detailed and then the advantages are discussed for using the SLRA are provided for both a fully libre company, as well as traditional companies.
2. The Problem: Proprietary Sponsored Research Agreements
The three primary aspects of PSRAs that limit innovation are: (1) oppressive default reporting requirements; (2) delays related to confidentiality and publication embargos; and (3) restrictive IP rules and licenses.
By definition, a researcher cannot be doing research when writing a report about the research. As academics are already heavily incentivized to document their useful results for peer-reviewed publications, any report writing is redundant and can be viewed as administrative waste [51
]. This waste is increased if the reporting requirements are more frequent. PSRAs are general documents that can reflect the worst case for reporting frequency.
For both academia and industry in highly-competitive fields, research is a race and any delays are counterproductive to the goals of the research collaboration. Delays that are part of a standard PSRA involve waiting times between when confidential information is disclosed orally or visually (e.g., in a presentation of video chat) and when it is converted to a tangible form marked “confidential”. Similarly, there is normally some form of embargo clause in a PSRA to ensure that the industry partner has time to read publications and presentations to ensure that no confidential information has been disclosed. These delays, which can add up to months, can represent a significant delay.
Finally, all PSRAs have language defining the use of IP, which can be complicated by which party is responsible for funding the patents, how licensing will be negotiated and reimbursed. In general, these clauses identify three types of IP: the company’s IP, the university’s IP and jointly developed IP. As libre companies do not use IP monopoly practices, by funding a project under a PSRA, they run the risk of enabling the production of IP on building block technologies [52
]. This same risk exists for companies that use IP conventionally. If the project is outside of an area they target for IP protection (and even if it is joint IP development) the PSRA clause can cause unnecessary legal costs. For example, if the university creates IP the business needs or they create joint IP that the university patents, the company could have technology paths closed off or be forced to pay to license technology they funded. This is obviously counterproductive to the goals of any company.
The differences between a PSRA and SLRA shown in Table 1
are substantial. First, the reporting requirements eliminates all unnecessary reporting, which otherwise slows research progress. Second, a clause is inserted in the confidentiality section (highlighted in red in Table 1
), which specifically attempts to minimize confidential information sharing and ensures that if any is shared that both parties are immediately aware of it. This reduces potential delays, which in standard PSRA contracts can be a month or more. Third, in the publication section the standard 30 days for review is shortened to 10 for journal, conference and other publications and is completely eliminated for patents (as there would not be any patents filed for inventions by either party related to the project in an SLRA). For libre projects, university and corporate partners would normally meet (even if virtually) at least weekly anyway—further delays have no benefit to either party. Rather than reducing the time to a week, 10 days accounts for someone being sick or away on a week-long vacation to ensure both parties can be compliant with the SLRA. Fourth, the intellectual property clauses have been significantly shortened as all of the inventions coming from the projects will be libre (again there will be no patents generated by the projects following and SLRA). Fifth, similarly the IP license section can be significantly reduced by detailing clear open licenses including a creative commons [57
] and GFDL [58
] for documentation, GFDL for FOSS and the CERN OHL [59
] for FOSH. These forms of licensing not only eliminate the negotiation period, license, and reimbursements for IP related legal work from the research agreement, but also eliminate the need for the costly legal work in practice. The aggregate of the effect of moving from a PSRA to an SLRA is that both companies and universities can reduce both time and costs for setting up collaborative research by virtually eliminating legal costs. In addition, university researchers can spend more time researching and providing value to the firm for the same investment expenditures (i.e., rather than spend time on bureaucratic waste and management). In the long term, the adoption of SLRAs could even reduce the overhead associated with university research, although further work is necessary in this area to determine what that reduction would be in percent of contract funding.
It is noteworthy that in most PSRAs in the IP section there is some mention that the sponsor recognizes that the University has an obligation to utilize the knowledge and technology generated by their own research in a manner which maximize societal benefit and economic development and which provides for the education of graduate and undergraduate students. There is substantial evidence that releasing innovations and inventions using open source licenses can indeed maximizes societal benefits [5
] as it leads to all of the benefits in FOSS [14
] and FOSH [30
] development discussed above. As the open source development of digital fabrication tools such as 3-D printers, mills and laser cutters [31
] are coupled with software that enables convenient customization [63
] and free public resources for open source design exchange [64
], peer production can emerge [66
]. These practices bring high value products to consumers for lower costs than what is available on the market for a wide range of consumer products [67
] ranging from toys [70
] and educational aids [71
] to upper end of scientific equipment [76
]. There is already considerable evidence that the downloaded substitution value [79
] of such designs (which provide savings from 90–99% [41
]) can bring significant return on investments (ROIs) [80
] to research funders. SLRAs could assist libre firms selling open hardware products for this market (e.g., Sparkfun, Aleph Objects, Arduino, re:3D, OpenPCR, OpenQCM, OpenTrons, Pax Instruments, etc.) accelerate their R&D efforts with universities. In addition, there is open source development of Internet of thing (IOT)-based energy monitors for buildings [81
], energy efficient homes and subsystems [82
], and even smart cities [83
], all of which could provide substantial ROI for government-based funders and their corporate partners working more easily with universities by using SLRAs.
However, it should also be pointed out that the same advantages can also be used in the business world following both libre and conventional IP strategies. There are a number of business models that can provide significant returns to firms using open hardware business models in this area [47
] including: (1) kit, fabricated equipment, and complete tool suppliers; (2) specialty component suppliers; and (3) calibration and validation-related as well as using FOSH/FOSS related services. All businesses that profit by using these models would also benefit from SLRAs. Finally, even firms with conventional approaches to IP may benefit from SLRA development. For example, larger companies and those with more experience and existing patents can use a “secondary supplier model” [84
] for open source sponsored research. This model works when the firm benefits from greatly expanding the size of a market (i.e., if they have a considerable market share). Then by sponsoring research at a university that would provide a new application for one of their existing products or expand the market directly by providing more users can provide high ROIs for the company. Using an SLRA speeds the process of collaboration with the university and allows more of the funds for the project to be focused on the project’s goals. For example, NECi, which is primarily a bio-tech firm that manufactures enzymes, sponsored the development of an open source photometer, which radically undercut the cost of other methods to detect nitrates with the use of their enzymes [85
]. By agreeing to release the designs of the tool under an open license, they encouraged new customers because their enzymes were necessary for the functioning of the device as designed. This invention effectively opened the market of lab-grade nitrate testing to the consumer and citizen scientist level, which had not been possible previously. Therefore, for example, they had established customers in the agricultural industry, but the device allowed the price point to be pushed down low enough to target small family-run farms and even gardeners.
There is considerable future work needed in this area of research agreements fostering innovation and invention. First, the benefits of the use of SLRAs should be quantified. For example, the costs for a range of businesses (both sizes and types) to run a PSRA through their legal counsel and negotiate with a range of types of universities to set up a research contract should be quantified. The research funds preserved by the aggressive time savings protections used in the SLRA could also be quantified (e.g., how many additional research hours are preserved by coupling reporting and publishing together). In addition, the impact of these time savings must be carefully evaluated as a benefit. Although time savings on bureaucratic functions may seem inherently beneficial, there is some evidence from the product development literature [86
] that there is a complicated relationship between the speed of product development and quality. Whereas research time spent on the unneeded functions of the PSRA may appear wasteful, these delays could result in some unintended benefits (e.g., longer time periods to test durability). Likewise, the research investment cost per invention under a PSRA and SLRA could be compared within similar research areas in a number of fields. The same output measure could also be adopted for comparing the peer-reviewed publications output from a given amount of research expenditures operating under a PSRA and SLRA. Finally, although some companies like Red Hat may have no issues with the “no patents” stance of the SLRA discussed here, other companies may not. Future research is needed to develop a sponsored research contract that could enforce a Red Hat-like patent promise [87
] or patent pledge [88
]. Therefore, any inventions developed from the sponsored research which were patented would adopt a clause similar to “any patents will be accompanied with an unrestricted grant for FOSS/FOSH” development. This approach would still have some of the benefits of the libre approach to technical innovation, but would lack all of the benefits from a full SLRA.