Along with a summary of the day’s discussions (which are also available to view online), the report contains a the authors’ synthesis of key principles and recommendations for preprint technology development (found in Table 3):
- Preprints are meant to facilitate and accelerate scholarly communication.
- Preprint services should encourage open science best practices.
- Meet researchers where they are now. Accommodate existing workflows and formats while moving toward best practices over time.
- Remember the motivations of researchers (including credit, career progression, and convenience).
- Take advantage of available technology. Preprint technology should be built quickly in a way that can be extended and expanded in the long term by many parties.
- Allow preprints to be transferred to journals in formats that fit journal workflows.
- Focus on standards. Use schema.org compatible meta-tags and recognized API standards such as OAI-PMH or equivalent. Use the standard persistent identifiers adopted by the community so that we can systematically link up resources, people, and organizations. For example, include person identifiers, document identifiers, identifiers for data, etc., and authenticate them to the extent possible.
- Make markup consistent. Engage with JATS4R or similar initiatives and follow existing recommendations on tagging.
- Develop open technologies. Permissive, open licenses on software should be strongly encouraged, and serve as the default for new code written for any ASAPbio projects.
- Encourage best practices for screening. Manuscripts must be screened by humans before posting, and takedown policies need to be implemented in a standardized fashion.
- Stay simple. Accept submissions in Word format and display them in PDF from day 1. The originally submitted files should also be retained and made accessible for mining and processing.
- Support open source conversions. Request and support the creation of an open-source document conversion tool from popular formats like Word and LaTeX to consistent markup (JATS and/or XHTML).
- Develop machine screening algorithms. To learn from the process, require all manuscripts (accepted and rejected) to be collected along with their screening status to form a database of content; use this to improve machine screening algorithms.
- Streamline transfers. Support simple transfer of articles to traditional journal workflows.
- Promote data sharing. The service should make it easy for authors to refer readers to data, software and other relevant materials. Encourage and facilitate deposition of data in appropriate repositories.
- Directly accommodate deposition of supplementary files (such as figures, movies, and text), which should be given their own unique identifiers and be preserved and indexed appropriately.
In accordance with the subject matter, this work has been made available prior to peer review, and we welcome comments and critiques.