How is software development like scientific research?

As a publisher of open access scientific journals, we know that the research world changes fast – assumptions that we make at the start of a project are likely to change while we are working. In our technology team, we work using a software development methodology that is built on similar principles to scientific research.

2

Imagine you have a research question to answer. How do you go about it? You come up with a hypothesis, set your parameters and define your research group. You collect and analyse the data. You adapt and evolve your research in response to the evidence you collect.

You wouldn’t dream of writing up your initial hypothesis and ignoring the evidence.

At BioMed Central, we think building a website is a bit like that – and that we can build better websites by using an evidence-based, user-oriented approach to software and product development.

So what does that mean in practice?

As a publisher of open access scientific journals, we know that the research world changes fast – assumptions that we make at the start of a project are likely to change while we are working. While we can take a pretty good guess and do thorough analysis upfront, we can’t predict 100% accurately what the research community will want and need in 18 months’ time.

Instead of guessing and hoping that we’re building websites that match how the research community want to read, find and use scientific content, we accept that things may change and we work in a nimble way that allows us to react and respond in a timely way.

In our technology team, we work using a software development methodology that is built on similar principles to scientific research. Researchers propose a hypothesis, analyse data, and develop a thesis; in technology, we build a little bit at a time, measure usage and user feedback, and use what we’ve learnt to refine and improve what we’re doing continually.

biomed central agile software development teams in action
BioMed Central Agile software development teams in action
Rachel Craven CC BY

Agile… what?

For me, Agile is about getting people to work together, listening to feedback and incrementally building software in a sustainable way. Lean moves to a more scientific and data-driven approach where we try experiments and record the results leading to a more iterative approach where even a single feature may be revisited multiple times as we learn more.


Dan Bodart
Principal Software Engineer at Springer Technology

Strictly speaking, at BioMed Central and Springer, we practice a hybrid methodology – a combination of Agile and Lean Start-up. Agile is a more people-oriented approach to software development, which emphasises collaboration over the more traditional process-driven approaches, like Waterfall, while Lean is a more data-driven approach emphasising iterative product development – making a new website or product piece by piece based on feedback, metrics and usage patterns, allowing for the process of evolving and learning that we like.

But why is this good for science publishing and science communications?

We believe using these methods of developing software will help the research community as scientific publishing evolves, ensure the most up-to-date and innovative ways of communicating science.

This approach gives us:

  • Quick feedback cycles – we can gather plenty of evidence to help in decision-making about what to build and why. Basing our choices on evidence means we can be more confident that the end product will be relevant and meet the needs of the community. For example, early in the beta process, lots of people told us via the feedback form that the font was too faint. This meant we were able fix it immediately, which is better for everyone!
  • Transparency – we love the way this methodology ties in with the BioMed Central ethos and the way of we like to work as an open access publisher.
  • Collaboration with the community – we think that the two-way conversation between publisher and research community means better, more relevant websites and author services.
  • Evidence-based investment in technology – we work by building the smallest piece, gathering feedback and analysing data, learning from and evolving with it before moving on to the next round. This means we don’t waste resources building things no one wants or needs.

Is that what the BioMed Central website beta is all about?

This is the approach we’re aiming to take with the rebuild of our open access journal websites. We’ve chosen to release our new websites iteratively as a beta to get as much feedback as possible during the development process. This means we can respond faster and more effectively, and build a better home for scientific content and journals.

You can view our work in progress and see what we’re working on as we’re working on it, as the beta sites are constantly evolving.  Every time we complete a feature, we put it live for people to try out and give us feedback. Then, we take that feedback and build on it, continuously developing our product in a user-centred, evidence-driven way.

How do I get involved?

Get involved in the conversation about where the technology powering scientific publishing is going- take a look at our work in progress and tell us what you think.

Interested in joining our Tech team? Learn more about us and see our vacancies.

 

View the latest posts on the Research in progress blog homepage

2 Comments

tushar

Hi, thank you for this post I agree with you that Agile is a more people-oriented approach to software development, which emphasizes collaboration over the more traditional process-driven approaches, like Waterfall, while Lean is a more data-driven approach emphasizing iterative product development. very useful information

Comments are closed.