Daniel Hook is CEO at Digital Science and in his free time continues to work in theoretical physics.
What did you train in? What are you doing now?
I spent 11 years studying physics and theoretical physics at Imperial College London. Originally, I joined the Physics with Theoretical Physics BSc program in 1996, I carried on to do a 1-year MSc in Quantum Fields and Fundamental Forces in 2000. I then studied part time for a PhD in Quantum Statistical Mechanics with Dorje Brody finishing in July 2007, submitting just before the RAE deadline. I’m now CEO of Digital Science, a technology company that aims to improve the research ecosystem by providing better tools for researchers, administrators, librarians, funders, publishers and corporates. While the leap from theoretical physics research to trying to improve how research is done is an improbable one, I will attempt to explain (below) how that happened.
How do you introduce yourself (I am a physicist/entrepreneur/…)
I always claim that Theoretical Physics is not a job that you do but rather it is the person that you are. As such, it’s difficult to answer this question since I’ve always felt I’m both physicist and entrepreneur – I certainly bring a lot of aspects of theoretical-physics thinking to how I approach business. Introducing myself as CEO, entrepreneur or academic all seem to be disingenuous to one or other of the communities of which I consider myself to be a part, so I usually introduce myself as “someone who helps software start-ups to support researchers”.
How did you your career progress from a PhD in theoretical physics to leading Digital Science?
That’s a long story, but an abbreviated version goes something like this. Carrying on in theoretical physics after a PhD usually means 5-10 years of postdocs in several geographic locations; the often-taken alternative being working for a bank as a quantitative analyst. Neither alternative seemed to be very attractive to me, or to my office mates at the time, so we founded a software company called Symplectic together. We liked academia, but had noticed that the software that academics had wasn’t too good, so we started working with a variety of parts of Imperial College to develop better software to support academics. In particular, the Faculty of Medicine was very collaborative and together we developed a piece of software that would later become Symplectic Elements, our research information management platform. By 2009, we had started to sell Elements outside Imperial College and had been noticed by Nature Publishing Group, who were already planning to launch Digital Science at the time. Symplectic became one of Digital Science’s first investments in 2010.
By 2013, I was spending about equal parts of my time working on Symplectic and helping to establish the Research Metrics group at Digital Science, which wasn’t really fair to either company. As a result, in the middle of the year, I moved to become Director of Research Metrics at Digital Science and Symplectic promoted Jonathan Breeze to become the new CEO of the company. Two years’ later, Digital Science’s founding Managing Director, Timo Hannay, decided to launch his own start-up SchoolDash and I was asked to lead Digital Science as his successor.
How did you co-found Symplectic? Do you have any advice for young scientists who would follow your career path?
Co-founding Symplectic, as I’ve mentioned, was in part a decision based on the idea that the four of us who co-founded the company didn’t want to leave academia, but also didn’t see a route to do theoretical physics in a way that worked for us. We also wanted to give back to an environment that we loved and where, through our PhD studies, we had seen lots of things that could have been done better with a good software solution. Luckily, in a lot of theoretical physics research, you usually need to learn some level of coding. In those early years between 2002 and 2008, the four of us wrote about 12 pieces of software from a web content management system to an examination management system. It was a great way to learn the tools of our trade and to learn how to run a company.
I would not recommend following my career path to anyone – it was very much a personal choice and one that, by luck, has turned out to suit me. That said, undergraduates and PhD students are often taught a definition of success that is very narrowly defined – specifically in the academic context. What I have learned from my non-standard path is that success can be many things and that ultimately it is about finding a way to make a difference in a way that is personal to you.
Why are you still involved in active research?
As I said earlier, I don’t believe that theoretical physics is just something that you do. I really love doing research and I’m very lucky in that the type of research problem that interests me is the type of problem for which I only really need a pen, some paper and perhaps a computer. At the same time, I happen to think that if you’re going to write tools for researchers you can only do that well if you understand what challenges researchers actually face on a day-to-day basis. As such, I think it’s important that I continue to do research to be constantly reminded of what the challenges are and what doesn’t work as well as it could.
I should also say that I’m very fortunate to work with some really great collaborators who put up with my very busy travel schedule and who continue to work with me after all these years.
What is your vision for the future of science communication?
This is a really complicated question. I’ve spent a lot of time thinking about this problem and I’ve given a few talks on it in the past couple of years. You can find one of them here. If you can’t sit through the whole 55 minutes of the video, then I can try to summarize my position as follows. I think that:
- Communication must become more open and more collaborative – I think that material will be shared earlier in the research process with a greater range of people and that there will be credit and incentives that help this to become a reality;
- The mechanisms that capture the research outputs of experiments or other data gathering activities will become smarter, more nuanced and more complete in the contextual data that they capture – current equipment and approaches are far too narrow and focused, and don’t capture nearly enough context around the experiment;
- Communication will become more iterative – we can already see this starting to happen in that researchers now release datasets independently of a publication; there are often versions to the dataset as more data are collected and added to the public release; preprints are also changing our relationship with versions of record and the concept of priority in research.
- We will move away from the scholarly article.
Ultimately, what makes the scholarly article and the monograph the two preferred forms of communication are three key factors: Firstly, the fact that they are published on a specific date. This allows them to, secondly, have a physical form, which happens to be fundamentally the same as one that we learn to interact with from a young age. Thirdly, that physical form encapsulates an elegant structure of information that quickly gives us contextual information about what we’re reading.
In short, we are conditioned to hold something in our hands that feels like a book. With research literature that is only possible because a particular version is published on a particular day. As Geoffrey Builder has observed, by just looking at the front page of a paper, any researcher can identify where the authors, affiliations, title, abstract, main text, journal name, page number, date and DOI are located in the layout without seeing even a single word. Indeed, in many cases researchers can identify the name of the journal from layout alone.
However, the past few years have seen the nature of research results in many fields change completely. An increasing number of researchers now have vast amounts of data that they need to share in order for their research to be reproducible; they have developed software; their data needs to be consumed as a video or audio file or using a specific viewer in order to interpret it. On top of this, many researchers are beginning to see significant value in sharing negative results to increase the efficiency of the research system. None of these aspects can easily be fitted into the standard, flat, paper-based article or monograph.
As a result, I see the principal research outputs becoming the research objects rather than the papers. I see a deep need to change research evaluation and incentives to take this shift into account. I see research communication becoming more like computer software in the sense that it should be highly versioned, highly collaborative and quite open. I believe that “co-authorship” of research objects will be fluid and changing in time. I think that research reviews may be created by AIs at our request – relating research objects that interest us and pulling together the thinking of multiple researchers to meet our current need for information.
Even if my predictions are not accurate, it seems clear that there are many opportunities to rethink how publication works and that there are a number of transitions that are likely to take place in the next few years.