Stavrina Dimosthenous from Data at Royce sat down with Adam Plowman to discuss his recent experience publishing in a fully open access journal, Materials Open Research, and MatFlow, a computational materials science workflow management Python package.
*You can find out more about Materials Open Research by following this link, and about MatFlow by looking at the MatFlow website. MatFlow supports designing and executing (on schedulers or locally) workflows that integrate with various materials codes, including the crystal plasticity code DAMASK, the crystallographic texture Matlab toolbox MTEX, the EBSD and DIC processing code DefDAP, and codes for material formability analyses.
Hi Adam, can you tell me what motivated you to publish open access?
An open peer review process seemed quite attractive to me. The novelty in our manuscript was the way we chained together several simulation and processing steps into a cohesive workflow, and we were looking for constructive feedback on our methods.
What influenced your choice of journal?
My PI heard about [Materials Open Research] and thought it would be a good idea, aligned with the ethos of MatFlow being fully open access and open-source.
So the open access model? How would you compare your experience with other journals?
My experience was good, I would publish with this open access model again. My initial impressions were that the editor check was quick and my paper was online and public (with a DOI) quite rapidly. So I could then share it, even though it was not yet in its final state. However, we did struggle initially to attract a sufficient number (three) of reviewers. It was around seven months into submission that we received our third review.
Why do you think you only just got a 3rd reviewer, seven months into submission
I can speculate a few reasons, although this is only speculation. It might be because [Materials Open Research] is a new journal, so many potential reviewers have not yet heard of it. Although, the suggested reviewers had an immediate and positive response on agreeing to review. It was a long process to get a 3rd reviewer. It might also be because our paper focusses strongly on method development, which some potential reviews may not find as interesting, even though it’s good work, and serves as a solid foundation for future studies!
What are your impressions of the open review process?
The open review process has been great for me. I think it acts as an incentive for reviewers to do a thorough job, since everything, including reviewer comments, will eventually become open access with the paper. In the interests of scientific rigour, publishing with this model seems to be a more logical approach. Since the whole process is open, a reader can track the changes made due to reviews.
So, do you think all reviewers read the whole paper?
Yes, I think if someone is reviewing a paper in a fully open review process, it is more likely they will read the paper carefully! We got very useful, constructive feedback and the structure of that feedback was not any different from the closed review process.
How was your correspondence with the editor?
They were very proactive and keen to keep the publishing process moving.
Overall?
I definitely prefer the fully open publishing model. Especially if you take the view that science is collaborative, the open publishing model can be a good tool to help you get the most useful and constructive feedback. The submission user interface was easy to use, and I used their own LaTeX template. In the end, I think there’s a better chance that you’ll end up with a better piece of work, and I think that’s due to the openness.
MatFlow
You recently (September 2023) held a workshop for MatFlow, the Python package you developed with your collaborators.
Yes, it was a full-day workshop held at the Royce Institute, with attendees joining both in person and remotely from several different organisations and research institutes.
What was your motivation for having the workshop?
MatFlow became an important output from the LightForm programme grant. We also thought it was important to hold a workshop that anyone (internal and external researchers) could come to, learn how to use MatFlow, and see how they could benefit from workflow management. I think the engagement with researchers will lead to lasting impact from that part of the work, it advertises the product, gets the work out, it cascades, and we might eventually get more contributors, all part of open-source projects.
Did you weave open-source principles into the workshop, then?
Open-source is at the heart of MatFlow. MatFlow is an open-source package, and when using MatFlow, you generate workflows that you can then inspect and easily share on public data repositories like Zenodo.
Do you have any interest from anyone wanting to develop MatFlow or add-ons?
I have had quite a bit of interest, yes. Someone is developing a Thermo-Calc module to MatFlow. A major focus of MatFlow now is that it is community developed and sustained.
Will you be continuing with MatFlow development in your new role?
I will be continuing to lead the development and maintainance of MatFlow in my new role at UKAEA, with a particular focus on developing machine learning and uncertainty quantification workflows.
Thanks Adam and all the best for your new role!
