On Monday I hosted a panel about how to create infrastructure that can support a trustworthy literature. I had an amazing set of panelists:
- Kaitlin Thaney, Executive Director of Invest in Open Infrastructure
- Hylke Koers from STM, and one of the leads of the STM integrity Hub
- Jimmy Wales, founder of Wikipedia
While I am sure that
I had a lot of compliments about the panel, and the depth of the conversation that we had. That was very rewarding. Though I have been presenting for many years I always get nervous before things like this. I had two helpful pieces of constructive feedback, and I had two thoughts after the panel about things I would have liked to improve on, but think we covered some very important topics.
Ahead of our panel we worked on an outline of what we wanted to cover, and we broadly followed it, allowing for a good amount of time for questions from the room.
This was the structure that I wanted to follow:
- 15:30 – 15:40 Super brief introductions and opening framing - As we think about the question of trustworthy literature, over the next 10 years what are the largest risks that you see that we have to design against (what would a bad outcome in 10 years time look like?) - Round robin getting about 2 – 3 minutes from each of you on this question to frame the rest of the discussion. What role do we see for infrastructure & standards to help reduce those threats?
- 15:40 – 15:50 How might you design infrastructures to be resilient (perhaps immune) to the ecosystem risks that we identified earlier in the panel discussion?
- What’s already out there? What is under development? Where are the gaps?
- Where might we take inspiration from, within and without scholarly publishing?
- 15:50 – 16:00 Infrastructure on its own is never enough, what are your experiences around designing for behavioural change, and what community structures have you found to be critical to drive adoption?
- 16:00 – 16:10 What patterns are there to support sustainability of infrastructure?
- Why is infrastructure hard, and sustainability hard?
- Do funding mechanisms create constraints? Is there something like a “reverse Conway’s law” that could be applied here? (https://www.thoughtworks.com/en-gb/insights/blog/customer-experience/inverse-conway-maneuver-product-development-teams](https://www.thoughtworks.com/en-gb/insights/blog/customer-experience/inverse-conway-maneuver-product-development-teams)).
- 16:10 – 16:20 Policy - given the focus of the meeting, what policy recommendations would be advocated for?
- Have we seen examples of successful policies?
- Equally are there examples of where policy decisions have had a negative impact ?
If you read through this you can see that the main message I wanted to speak to was about how the support of infrastructure is highly dependent on sustainability models, policy, people. We could have spent the time shoe gazing about what things AI might do in the future, but given the amount of times infrastructure was name checked across the two days (often in terms of infrastructure), I wanted to make sure that we emphasised the contingencies of infrastructure.
On utilities, remember utilities often have mandatory taxes which publishing does not have, and digital infrastructure is often so invisible that the true cost of maintaining it is often massively underestimated (https://www.fordfoundation.org/wp-content/uploads/2016/07/roads-and-bridges-the-unseen-labor-behind-our-digital-infrastructure.pdf).
I think we covered these points really well, and I was delighted with the contributions from the panel.
What I thought our panel brought to the discussion was an amazing diversity of experience around the creation of sustainable infrastructures, because that is at the heart of how you answer this question. If you want your infrastructure to support a trustworthy scholarly record, first and foremost that infrastructure needs to be sustainable. It is software, and by virtue of that “soft” part, it can change, but if you don’t have the people who can change it, it does not matter how clever it might be.
To be explicit about the make up of the panel, my view on this was that Hylke through his work with STM has a view into how infrastructures are built out and supported within a commercial publishing setting. Kaitlin has worked with many organisations in the open space, helping many of them move towards viable sustainable pathways, and not just dependent on grant funding, and Jimmy has worked on leveraging crowds to contribute to the creation of a sustainable model for Wikipedia. That alone shows us that there need not be one model.
The two criticisms I had that I thought were very fair were:
- The discussion seemed to be three parallel views, not connected. I think that was a reflection of the diverse views we brought together for the panel, and in fairness I heard the opposite view from others.
- Someone at dinner told me that they didn’t know who the people on the panel were, or what they did (apart from having heard of Jimmy). That was on me because I decided to go straight to discussion and leave folk with the very extensive information from the conference booklet to figure out who people were, on the basis that we only had an hour, but on reflection I could have allowed at least a small introduction.
My own thoughts after the panel were, on reflection, that I would have liked to have had:
- Some more very specific examples of how we had seen infrastructure work to support trustworthiness, e.g. the moderation system in wikipedia, the ability of Jupyter notebooks to be a baseline for reproducibility, the way STM integrity hub can identify duplicate submissions, and so forth. This was not something that I had thought about during the preparation of the panel, but I definitely felt some points like that could have been beneficial.
- I realise now it could have been awesome to get some real costs on the table, just to help with grounding the audience in a sense of this. In a world where Mark Zuckerberg is paying $200M for one engineer it does not hurt to point out that the skills needed to run our infrastructures are both under high demand, and costly.
In the end though, I think it was a valuable contribution to what I think will turn out to be an important meeting, and I remain really grateful to the speakers, and to the opportunity to be part of that event.