Looking Back at the First Quarter of Twenty-First-Century Standards

Letter from the Executive Editor, January 2026

We often think of digital technology adoption as occurring at a breakneck pace. If you’re not the one to implement or adopt the latest tech trend, you’re woefully behind the times and risk never being able to catch up. The reality is that implementation is somewhat slower, more measured, and less radical in the short term, despite the hype. Let’s start off the new year thinking longer term.

Take a moment and consider the closing of the first quarter century and how standards have influenced the many changes in our community as well as what portends our future. Digital publishing was in its very infancy at the dawn of the second millennium, and libraries were primarily focused on managing print collections. At the time, both libraries and publishers looked not terribly dissimilar from how they had looked for decades, even centuries. However, people could see how new tools and approaches were gaining traction and were well on the way to begin implementing a new digital future. Many publishers had implemented internal digital production processes and begun moving rapidly toward digital formats for journals, although e-book adoption—despite the excitement—was still quite nascent.

According to the Association for Research Libraries' LibQUAL statistics, in 1999–2000 the median ARL expense on e-serials was only $735,317 and the median expenditure for all e-resources $931,210. At the time this represented just 12.8% of the total materials expenditure. Across all ARL libraries that participated in the survey, total spending on electronic resources was $95 million, whereas the total materials expenditure was more than $750 million. The digital revolution had begun, but it was still in its infancy.

File formats and content models to create these electronic documents were often bespoke. The most common online format was PDF, and hard-coded HTML was the only other option. Most publishers had custom-built content production pipelines or were in the process of defining them. This inhibited not only data exchange and content interoperability, but also things like long-term preservation. In 2002, work on the Document Type Definition (DTD) specification started at the National Library of Medicine (NLM) for its nascent PubMed Central system. This work evolved from earlier publisher-specific DTDs and provided a common XML format for journal article exchange and archiving. It would eventually become the basis for the NISO Z39.96 Journal Article Tag Suite (JATS), standardizing academic publishing content for data exchange and preservation. By providing a consistent structure for research outputs, JATS allowed for content to be structured both for human and machine consumption. It also became a lingua franca for content exchange and validation that supported better discovery and long-term preservation. The model eventually became the basis of other content models for books and standards and the basis of the Manuscript Exchange Common Approach (MECA). It serves as the vehicle for many other NISO standards, such as CRediT, article version information, Access and License Indicator information, and more. And it fostered the development of many startup tools and resources, because it allowed for content ingestion from most publishers without having to retool for every new content provider.

As the pace of acquisitions of digital content grew, managing these resources with systems designed for print materials quickly became unrealistic, but there were no commercially available systems for separately tracking digital resources at the start of millennium. Most often, locally developed databases were created to handle the complexities of electronic journals, which were not well managed by traditional Integrated Library Systems. The need for standards around data exchange, terminology, and interoperability of digital resources also became obvious. In 2002, NISO and the Digital Library Federation sponsored a Workshop on Standards for Electronic Resource Management, which kicked of DLF’s Electronic Resource Management Initiative (ERMI) effort to standardize these early, fragmented efforts. NISO supported further development on ERM systems with the ERMI2 effort and several standards initiatives that developed out of this, such as ONIX for Publication Licenses (ONIX-PL), the Standardized Usage Statistics Harvesting Initiative (SUSHI), and the Knowledge Bases and Related Tools (KBART) Recommended Practice. These projects all attempted to streamline the challenging business of managing information about digital resources. While not every effort was a success, many of them have become core elements of the ecosystem for managing digital resources, including tracking usage data with COUNTER Metrics, TRANSFER, resolution systems based on OpenURL, federated access control through SeamlessAccess, and entitlement information using KBART. Without these standards, managing the hundreds of thousands of resources would be prohibitively time consuming, if not impossible.

With this scale, other problems became apparent. ARL published its first directory of e-journals and newsletters in 1991, and 110 titles were listed. By its final edition in 1999, it listed thousands of titles. Discovery systems for all this digital content also expanded. This expansion has included the rise of indexed discovery services and the launch of Google Scholar, Scopus, EDS, Primo, WorldCat, and PubMed and other search tools. From our initial work on metasearch and Open Discovery to potential projects related to Generative Artificial Intelligence and Web-Scale Discovery, NISO has sought to support efficient and accurate content discovery. A new generation of AI-powered tools is entering the marketplace, such as Semantic Scholar, Consensus, and Dimensions AI. These tools face the same questions that indexed discovery tools faced 15 years ago: what is included, how content is ranked, and whether these systems are free from bias. 

Looking forward to the next 25 years, a lot of work remains to complete the scholarly research ecosystem, to make data exchange more seamless, to populate it with specific identifiers and metadata extensions, and to facilitate content creation, discovery, reuse, and preservation. Tracking people and objects across the network to ensure their validity, such as online identity management for validation of authorship and recognition, content provenance, and access control, remain significant issues. Defining new content models for emerging methods of sharing research outputs, like virtual research environment interactions, visualizations, and software packages, will be an ongoing effort. These new along with other existing types of research outputs or elements of the research ecosystem should be either identified and described or—as with developing systems—implemented in the scholarly communications ecosystem to connect all the pieces in the scholarly research graph. We have only just begun to grapple with the implications of AI tools for publishing and their appropriate application for communication. 

These projects and initiatives are certainly not the only work accomplished in the past 25 years, and I’m sure many of you will have other thoughts on your favorite standards project from the past quarter century. Many of these areas are still works in progress, although some initial advances have been made. Others will require ongoing maintenance and support to remain relevant and fit for purpose. We certainly have our work cut out for us for the next two decades, at least!

Happy New Year, everyone!  May the next quarter century be peaceful and prosperous for all of you.