Publishing for reproducibility: collaborative input and networked output
Coko has a broad mission to reformulate and improve how knowledge is produced and shared. As a starting point, we’re focusing on scholarly publishing because, while the knowledge is crucial to us, the process is convoluted, expensive and slow. It can take months to even years for finished research to be published and the final product is missing key components such as the data, protocols, code and materials needed for other scholars to reproduce the work.
Evolving the scholarly communication workflow
Most scholarly communication workflows are still very much as they were 10 or 15 years ago, with work being done in isolated, proprietary tools and formats and these being shuffled around between people using email.
Because the very process of producing scholarly knowledge is flawed, the final published product, be it a journal article or a book chapter, is severely limited. At each stage, the research is dumbed down, data and context removed. To improve the published product, we must add value at each stage rather than remove it.
Coko is a community effort to build collaborative processes and technologies that respect the research process and improve the output. The open source tool chain being assembled enables real-time collaboration around the content and data as the published record is being produced. This means that the published product can mirror or exceed the research itself.
Shifting from a linear, largely offline submission and editorial workflow to a collaborative webspace means taking creating a digital first process that posits an HTML document at the center of a flexible set of tasks and action. The Coko community is building many tools including a sophisticated web editor and a workflow engine that can be configured for many different workflows and be adjusted easily to fit changing process needs.
Creating a community of exceptional open source projects
Coko’s philosophy is to bring together the best open source projects into an interoperable tool chain rather than try and build everything as one monolithic platform. For example, The Coko community is working closely with Substance.io which has built the first open source real-time and concurent web editor focused on scientific and scholarly content. All of the work is done within the document itself and using web collaboration tools for discussion and annotation. All actions are tracked to make the process transparent
And Coko and Substance are partnering with Stencila, a platform for creating documents that are driven by data, enabling templates for embedding data analysis and presentation code that work well within Substance’s WYSIWYG editing tools.
HTML first: content conversion and adding intelligence
Authors overwhelmingly prefer to write in MS Word. Rather than try to change author behavior overnight, it’s possible to transform Word and other largely unstructured documents into highly structured HTML early on in the production process, such as upon submission to a publisher. INK is another tool the Coko community is working on for content converters and other transformation tools and acts as a job management system, allowing users to configure a recipe of actions upon a document as it is being converted into HTML, such as enriching with identifiers, entities, links and semantics. INK normalizes metadata and connects all of the research objects that went into creating a body of research work, laying the foundation for a networked publication that will improve reproducibility.
The networked, living publication
By building the publishing process directly from the research objects such as data, code, materials, media and discussion, the collection of Coko community technologies will improve the published output, enabling a dynamic and evolving body of work to replace the static journal article or book chapter of today.
Post by Kristen Ratan, small improvements by Adam Hyde