Success at Scale

In support of our mission to accelerate progress in science and medicine through innovations in the way research is published and communicated to a wide variety of stakeholders, PLOS has either built or integrated third-party applications, looking to the best available tools. Our journal websites have been powered by a custom-built application called Ambra™ and we have used vendor-supplied applications such as AllenTrack and Editorial Manager for our submission, editorial and peer review operations.

As Open Access increasingly becomes the publishing option of choice for authors, and is increasingly mandated by funders from private foundations to governments, PLOS experienced the challenge of success when manuscript volume began to outpace our technological capabilities. The submission and journal software systems that helped to support our growth and evolving needs over the past decade were increasingly seen as hindrances for rapid innovation moving forward.

Several years ago, we forecast a need to enhance and implement sophisticated content management systems for optimum handling of web and article content, as well as to innovate architectural advances to the journal website software. In 2013, we completed a landscape study to determine the available options for addressing the concerns with our current application suite and as a result, began the challenging effort of building our own submission, editorial and peer review system.  We also made a decision to modernize Ambra into a Service-Oriented Architecture (SOA) application to enable us to more rapidly deliver new features to our readers.

People First

The first thing we did was modify both our organizational structure and culture around software development, approaching each of these endeavors using Agile methods. This required a move to a project-based organization structure and involved a great deal of formal training and coaching for the dedicated teams. We also committed to the “incremental value delivery” mantra of Agile, breaking our large objectives into smaller deliverables that could actually be tested by the various stakeholders as we moved forward. Each of our development teams work in two-week sprint cycles that enable testing of delivered components over relatively short intervals. We launched all of these efforts in earnest at the beginning of 2014.

We then turned to the needs of stakeholders and the manuscript submission system, the core of what drives PLOS’ ability to bring scientific research to the community as rapidly as possible without compromising on quality and integrity. During the first 18 months of development, we focused primarily on the overall architecture and author submission process.  Our goals were simple:

• Focus on the needs of our users, placing authors, reviewers and editors at the center of our development efforts
• Reduce the amount of time and effort authors need to submit a manuscript
• Streamline the review process to promote faster decision making
• Create an environment for all communication and collaboration needed throughout the lifecycle of a manuscript
• Deliver a modern, intuitive, browser-based interface and user experience
• Build a foundation that allows for rapid modification, customization, experimentation and future development

A Framework for Innovation

We worked on delivering an overall proof of concept that demonstrated our approach and supporting architecture to achieve our goals. Instead of diving deeply into any particular feature set, we delivered a minimum set that enabled us to test drive the initial concept and adjust our design as needed. The use of Agile methods was instrumental in allowing us to learn what worked and what didn’t with actual user interactions. Many of our initial concepts were completely overhauled based on these learnings and we have continued to respect this process during all stages of development. In the summer of 2015, we designed an overall architecture using a Ruby/Ember/PostgreSQL technology stack that is loosely based on the following components:

• Core Workflow Framework: foundational structure that supports the setup and control of tasks/cards
• Custom Cards: generic and custom forms/tasks and/or applications that can plug in to the workflow framework
• Ingestion/Export: module that supports the ingestion and export of documents and objects
• Event Server: module that logs all changes and events within the system

The combination of Core Workflow Framework (Core) and Custom Cards (Cards) has allowed us to balance the need to centralize core functions and features for modularization and cloning across the system while allowing flexibility for custom elements and processes to support future innovation.

Once the Core was sufficiently developed, we began to exercise! We developed Cards for author submission, reviewer, publishing and editorial staff. While we are one PLOS, our journals have unique strengths, missions and contributor communities. We wanted to be respectful of that and develop a flexible system to serve all stakeholders. Throughout the process and to this day, we involve end users in live testing, iterate based on feedback and strive for bi-weekly code releases. Our Minimum Viable Product (MVP) approach led us to focus on the needs of PLOS Biology, one of the smaller volume journals within the PLOS suite of journals.

Moving Forward

All of the custom cards and tasks necessary to support transition of PLOS Biology are complete. Subsequent journal launches will follow a rapid cloning process where most cards can be reused and customized workflows can be accommodated by the flexible and user-friendly architecture. We are implementing the appropriate IT infrastructure to adequately support a production instance of the application, while continuing to develop the system in support of our other journals. We have also built a Named Entity Database (NED) for individuals and institutions to source accurate identification and attribution across all PLOS applications and processes, including a revamped registration and user management module that was recently launched.

The best way to demonstrate the benefits of our new system is to showcase it on our own journals and communities first. Ultimately PLOS is committed to share what we have developed with the scientific publishing community and to work with others to enable Open Access and Open Science to flourish. PLOS’ Article-Level Metrics made available to the community under an Open Source MIT license – and now called Lagotto – have since been adopted by CrossRef. After launching on several journals to ensure scalability, the code base for the submission system will be made available via an open source MIT license so others can benefit from our efforts. We continue to seek interested parties to join our development efforts so that we can learn how to best govern the code in a fully open source environment.

To honor our roots in the Open Access movement while harnessing technology to propel scientific communication into the Open Science era, we chose the name Aperta™ for our new submission system. Aperta means “Open” in Italian, and brings with it the association of forthcoming and fairness, qualities that PLOS strives to bring to the process of publishing scientific research.

Learn about Aperta in the broader context of recent progress at PLOS in bringing authors to the center of science communication on The Official PLOS Blog.

Image Credit: Gerd Altman, Pixabay.com