VII · 46

Related Work

OpenURL has not been replaced. Adjacent work — IOTA on link quality, KBART on knowledge bases, direct linking through DOI — shapes the environment in which it continues to operate.

NISO IOTA

The Improving OpenURLs Through Analytics (IOTA) project, supported by NISO and concluded in 2013, set out to measure the quality of OpenURL traffic in production. By analysing very large samples of real OpenURL requests submitted to participating resolvers, the project produced metrics on what fields were typically present, how often essential descriptors were missing, and which source systems were producing the highest- and lowest-quality OpenURLs.

IOTA's findings informed work on emitter quality and provided the empirical foundation for several recommendations carried forward into KBART and link-resolver implementation guidance. IOTA itself wound down after its final report; the project did not result in a successor standard.

KBART

The Knowledge Bases And Related Tools recommended practice (NISO RP-9), discussed in detail at OpenURL & KBART, took up a complementary problem: not the quality of the OpenURL itself but the quality of the data in the resolver's knowledge base. Improving KBART data improves the accuracy of OpenURL resolution without changing the OpenURL.

Direct linking

The wider scholarly-publishing environment has moved toward identifier-first linking. DOI links are stable, simple, and increasingly accepted as the canonical link form for an article. Article-level identifiers in arXiv, PubMed, and similar repositories follow the same pattern. Where every reference can be resolved to a DOI, the question of "which copy" reduces in part to the question of where the DOI proxy resolves to — though even then, the institutional question of access remains, and so does the role of the link resolver.

Why OpenURL is not being replaced

OpenURL continues in production for three reasons that none of the adjacent work fully addresses:

  • Citations without identifiers. Many citations — older articles, books, dissertations, grey literature — have no DOI. OpenURL's descriptive form remains the only standardised way to carry them.
  • Institutional context. DOI resolution alone cannot answer "which copy is licensed for this user," because identifiers do not encode subscription state. OpenURL's separation of identity from delivery is the architectural feature that the appropriate-copy problem actually requires.
  • Installed base. Every major link-resolver product accepts OpenURL; every major bibliographic database emits it. The cost of switching out a working, federation-wide protocol is high.

There is no work currently in the NISO pipeline aimed at replacing Z39.88-2004. The standard was reaffirmed in 2010 and remains current. The Registry has been frozen since 2022. OpenURL is mature: completed, stable, and quietly carrying very large volumes of scholarly linking traffic.

Sources