Enabling OSS Innovation in Universities

Article by Aoife Tierney, Jiří Marek, Michael Nolan, Noel O'Connor, Sayeed Choudhury, Colleen Maloney, Clare Dillon and Ciara Flanagan

Society has reached a tipping point in terms of the global use of open source. The benefits of open source development and research is widely recognized and this is underscored by the recent introduction of policies across public and private institutions and governments. However, the social norms and infrastructure needed for capacity building and incentivizing OSS innovation and open science in the university context are in a nascent phase compared to the traditional model of research development.

The benefits of open source and open science from universities will only be available if it is properly encouraged and supported at the university level. But what does that look like and where do we need to focus?

The recent OSPO++ conference on ‘Open Source Innovation in Universities’ brought together an accomplished panel to share their experiences of enabling open source innovation in universities. Aoife Tierney, IP Development Manager at Trinity College Dublin; Jiří Marek, Open Science Manager at Masaryk University; Michael Nolan, Associate Director at Open@RIT; Prof. Noel O’Connor, CEO at Insight SFI Research Centre for Data Analytics; and Sayeed Choudhury, Director of the Open Source Programs Office at Carnegie Mellon Libraries discussed key innovations emerging from universities, best practices for supporting such research and more.

Enabling open source and open science within universities

Despite widespread agreement that open source research and projects are essential for future innovation, the systems to enable open source and open science are often lacking in universities. Creating Open Source Program Offices (OSPOs) is key to facilitating open source, and fortunately, universities now have examples of successful OSPOs to take lessons from.

Choudhury referenced his experience of building the OSPO at Carnegie Mellon University as an institutional effort to encourage people to become champions of open source and open science within the university. He recommended that university OSPOs be designed to empower researchers by providing the infrastructure, guidance and support that can be difficult to find elsewhere.

“We have the coalition of the willing and people in the church but they still need help and they still need support,” Choudhury said. “Faculty at US universities are much more eager and willing to take help when it comes to open source software than they are with data and articles, where they have a much better sense of what to do.”

Jiří Marek echoed the view that university offices can provide essential help and guidance - not just for faculty and researchers but also for project managers. He gave an example from Masaryk University, where OS project managers were assisted to write proposals. It was an opportunity for people with expertise in the open source domain to share knowledge and to empower others for success.

“It’s like scaling up to the general public because we have to consider open source, open data - everything,” Marek said. “It’s about the open principle – that horizontal principle – and we sometimes think about it as a vertical one. In my opinion, that’s wrong.”

Facilitating intra-university OS collaboration

Existing university OSPOs and the experience gained in their creation offer a number of best practices to guide future development. One important benefit of this could be to set basic rules of engagement around open source in universities, or as some panelists said, to “put some manners on it.”

The panelists agreed generally that open source software is often released from projects within universities in an ad hoc way that allows for limited, if any, control. OSPOs can help researchers and institutions understand that open source provides for more options than they might think.

Tierney cited an example of a project from Trinity Biomedical Sciences Institute that provided a statistical framework for analyzing dense sets of data. Because the work was open sourced correctly, it has enabled multiple other projects. “We were able to release that open source,” she explained. “And in parallel, there are other innovations coming out of that same research project that maybe would be more appropriate for patenting or for closed protection.”

Tierney also spoke about educating the research community collectively within Trinity on using and bringing open source into their projects. “There are certain disciplines like natural language processing or computer science that are more familiar with open source,” she said. “But we’re seeing that organizations like Horizon Europe are pushing for interdisciplinary research, so we’re bringing together people from traditional fields, for example, life sciences or biology or engineering, and [we’re] supporting them to work collaboratively. Open source and open policy facilitate that.”

As interdisciplinary collaboration gains traction, it will open more doors for collaborations that bridge the gap between researchers, institutions and funders. However, success hinges on all parties involved gaining better understanding and clarity on releasing and licensing open source.

An example of this came up when Choudhury referenced a project he was closely involved in with John Hopkins University’s OSPO and the city of Paris on an open source municipal service platform. Working with a local community center, the team built a demonstration of how the center could use the platform, following the allowances in the open source license. When the legal questions came, Choudhury pointed to the license.

“When various legal and administrative offices of Johns Hopkins asked, what’s the agreement for this? What’s the grant? What’s the data use agreement? I said, “There aren’t any, it’s the license.”

Best practices around university-industry collaboration

Many of the same considerations come into play when looking at how universities can collaborate with industry on open source research or OS-powered projects. When questions were raised about incentivizing startups to collaborate with universities, Nolan suggested the OSPO or tech transfer office within a university would be the right place to start making connections of that kind.

Nolan stated that for industry and university researchers to connect, “The work needs to be discoverable by the types of companies that are searching for it. It needs to be well documented. There needs to be a process for when the company wants to contribute back and further that knowledge. Then you need processes for making sure the faculty generating the knowledge is able to use that as a case for tenure and promotion.”

“Supporting these activities is where we see impact that is useful to society at large,” Nolan continued. “But also useful for the university to show that the research they’re creating has real-world impact.”

O’Connor offered an excellent example of university-industry collaboration, using the open source project, Saffron, developed by professors Paul Buitelaar and John McCrae in Insight at the University of Galway. The project is a knowledge extraction framework that had been used across multiple open source and public projects primarily in Europe, when it was used by the multinational corporation, Fidelity Investments, in a collaborative research project around better customer support.

“Saffron was basically the calling card, if you like, for us as a research center to be able to convince Fidelity Investments of our capability,” O’Connor explained. “It showed there’s a long track record here and deep expertise here, which led to multiple engagements and repeat business.

Addressing challenges to global collaboration

Despite the benefits and accomplishments evident in the global open source community, several issues remain with respect to funding OS projects across international and regional divides. The end goal as outlined by the panel is to continue to educate funding agencies and government departments on the benefits of open source and open science and the merits of core funding to provide sustainable support for researchers and projects in this field.

Marek talked about the need for a policy vehicle to move the needle further toward collaboration and knowledge sharing from within academia across institutions and national boundaries. One suggestion was to extend the Interoperable Europe Act beyond the public sector to include academia as well. Marek also cited the Brno Declaration as an example, noting it could serve as a template or be endorsed by the community rather than needing to start from the beginning.

Changing the metrics and not just the conversation

To conclude the discussion, the panel discussed the need to measure impact and innovation as a way to demonstrate value. The big question on everyone’s mind was how to change and view the metrics within the university context.

Nolan explained: “The challenge really is to help develop our institutions that are producing research in order to encourage them to measure [open source and open science innovation] and to evaluate it in a way that people can be rewarded for their work.”

Everyone agreed that metrics are good as long as they drive the right behavior, suggesting that further alignment and agreement is needed on how work can be appraised.

One point everyone agreed on was that while the reasons for developing open source innovation in universities are varied and important; OSPOs are a key enabler and therefore, are essential to its evolution over the years to come.

View the complete OSPO++ panel discussion here.

This work is licensed under a Creative Commons Attribution 4.0 International License (CC BY 4.0)