scope:

Introduction

These NISO/ALPSP Journal Article Versions (JAV) Technical Working Group recommendations provide a simple, practical way of describing the versions of scholarly journal articles that typically appear online before, during, and after formal journal publication.

Researchers, their institutions, and journal publishers are rapidly moving on from using static, single copies of research papers that are essentially "images" of a printed document. Changes in the way we create, edit, circulate, validate, publish, find, use, and update articles are producing multiple versions whose status and provenance is often unclear. Online searching now allows multiple versions to be found but rarely makes clear the relationships between them.

This is not simply a problem for the journals themselves. Repositories want to provide authoritative records of their faculty's work; libraries want to offer "appropriate copies" to different users; readers need to know what has been peer reviewed; and authors may wish to update their work and ensure that the latest version is used.

In September 2005, NISO (the National Information Standards Organization) launched a partnership with ALPSP (the Association of Learned and Professional Society Publishers) to bring together experts from the publishing, library, library systems, and user communities to examine the problems associated with the proliferation of different article versions. A Technical Working Group produced the recommendations in this report, with feedback and advice provided by a wider Review Group (members are listed above).

The recommendations have been agreed by all members of the Technical Working Group. The Review Group provided many points of feedback, advice, and criticism. This process was very valuable and led to improvements throughout the document. It was not possible to obtain a consensus from all members on all issues. However, all the proposals, comments, and responses are available to read on the NISO website and the comments received during the formal review period are incorporated here in Appendix 3 together with the responses to them.

The Technical Working Group created use cases to explore the lifecycle of journal articles, starting from a base case that describes a typical interaction between author, institutional repository, and publisher. Analyzing these led us to identify common lifecycle stages, the dimensions that describe the evolution of articles, and possible attributes of each instance of an article version.

Considering attributes helped us to agree upon the most important variables for describing versions. Ownership, bibliographic context, identifiers (e.g., DOI), relationships, fixity, and peer review are explicitly stated in the terms and definitions that we recommend. They can be described by the article version names and some are already covered by standard metadata elements (e.g., bibliographic reference, date, DOI). Other variables describe digital copies or variants, but are less closely related to the academic content of the article: visibility (limited, general), version status (known/unknown), source (different websites), scope (text, fullyfeatured, resolution of images, etc.), delivery context, and format (PDF, HTML, etc.). These are important factors and could be described in article metadata; however, in most cases it is possible to distinguish properties inherent to the article version and properties derived from the delivery system. Our focus has been to describe which version of the intellectual content of the article a reader has encountered.

These variables led us to identify important dimensions for an article version:

• Time: from first draft to latest version

• Added value: from rough draft to polished publication

• Manifestation/Rendition: different document formats and layouts

• Siblings: multiple mappings between technical reports, conference papers, lectures, journal articles, review articles, etc.

• Stakeholders: author, editor, referee, publisher, librarian, reader, funder

Manifestation was largely excluded from the nomenclature as being too dependent upon technology changes and the creator or user's computer systems (though it is relevant for describing our Enhanced Version of Record). Siblings were discussed at length and eventually excluded, as the relationships are complex and conventions vary considerably between disciplines. Our recommendations could be used to describe, for example, a technical report validated by a funding agency and published in an institutional repository, but we stuck strictly to our brief of considering only journal articles as a first goal.

The variables, dimensions, and use cases reveal the difficulty of describing all possible lifecycle stages in clear terminology. However, the vast majority of journal articles do pass through the same milestones and lie within a limited range of use cases. This remains true even though different users will value different versions for different purposes. From these common milestones we have produced six terms to describe journal article versions. The variety of activity illustrated in the 26 use cases (Appendix 2) shows just how much can be described by these high-level semantic terms. We have focused on changes to content, the formal social process of validation, and the ways in which journal articles are used. We have also noted the tradition that journal articles record the "minutes of science" and are intended as a fixed record of a body of work at a moment in time chosen by the scholar. This leads us to the Version of Record as a useful definition for formalizing academic achievement, distributing authoritative information, and building upon the established knowledge in a field.

In line with the JAV Technical Working Group's charge, the group submits the following:

• Background and Rationale: a narrative that explains the background to our project and the rationale for our recommended terms and definitions

• Recommended Terms and Definitions for Journal Article Versions

• Appendix 1 - Graphical Representation of Journal Article Versions and Relationships with Formal and Gray Literature; Assumptions, Primary Challenges, and Best Practices

• Appendix 2 - Use Cases: a set of use cases showing the application of these recommended terms

• Appendix 3 - JAV Review Group Comments: comments received from the JAV Review Group to an earlier Technical Working Group document submission, and the Technical Working Group's responses

We propose that the terms as defined be promulgated by NISO/ALPSP to the full journal article stakeholder community (authors, readers, libraries, publishers, aggregators, archives, repositories, research institutions, funding agencies, and service providers such as search engines and link resolvers). The JAVTWG recognizes that adopting a standard terminology will not be enough; to avoid version confusion, terminology needs to be implemented in such a way that readers (whether human or machine) encountering any version can immediately ascertain which it is and know whether it is trustworthy.