Release Notes

What's New

Recent platform and methodology updates in openscience.works.

Last updated: 2026-06-17

June 2026 · Citations & explorer

Merged DOI version cards

Versioned DOI variants (preprint → journal, v1 → v2, etc.) now roll up into a single explorer card with aggregated citation totals. The browse index no longer lists every version as a separate story.

version rollupdeduped indexcitation totals

Per-version citation breakdown

Article pages with multiple DOI versions show a version citation breakdown table in Impact metrics — each row links to the DOI and its citation count, with preprint lines included where applicable.

ArticleStorypreprinttransparency

Smarter site search

The explorer search bar now matches DOIs, author names (from all schema fields), and multi-word queries. Author profile pages are included in the haystack so you can jump from a name to their stories.

DOI searchauthor indextokenized match

June 2026 · Repository stewardship

re3data repository profiles

DataStory and SoftwareStory pages now surface stewardship context from the re3data registry: repository name, access type, certificates, PID systems, and preservation policies. Helps evaluate trust and reuse conditions at a glance.

Where this appears
  • Scientific Stewardship section on DataStory pages for the deposit repository.
  • Linked datasets on ArticleStory and BookStory pages inherit registry metadata from their deposit DOI.
  • Hero badges call out certified or open-access repositories where applicable.
re3dataDataStorySoftwareStorystewardship

June 2026 · News, courses & signal quality

News coverage on all story types

Google News matches from SerpAPI now appear in the News section on every story page, and news hits are indexed in the explorer. Previously some enriched stories did not surface their coverage.

news_coverageSerpAPIexplorer index

URL-backed news matching

News enrichment now validates article URLs against the work's title and authors before promoting a hit, reducing false positives from homonymous surnames and generic title words.

relevance filterURL validation

OCW index for teaching signals

OpenCourseWare and syllabus evidence is backed by a local index of MIT OCW, Open Yale, and other global course catalogs. Pedagogical Anchor assignments draw on indexed course pages instead of repeat live searches.

OCWPedagogical Anchorcourse catalogs

June 2026 · Story page UX

Section navigation from visible content

Long story pages now have a sticky section nav built only from sections that actually render — no dead links to empty blocks. Award lists and other long tables fold behind expand controls.

sticky navfoldouts

AuthorStory links

Author names in the hero line and contributor lists link to AuthorStory profile pages when a matching profile exists on the site.

AuthorStorycross-linking

LibraryThing reception cleanup

Reader reception from LibraryThing now filters malformed descriptions and shows ratings and reviews only when the platform returns usable enriched data.

LibraryThingBookStoryreader reception

June 2026 · Web mentions & public discourse

Science podcast mentions

Story pages now surface podcast episodes that reference a work, drawn from curated pilot sources. Each mention links to the episode, source show, and discovery date — visible under Public discourse on individual stories.

web_mentionspilot sourcespodcasts

Blogs & newsletters alongside social signals

Expert blogs and Substack posts join Crossref Event Data news/blog events in the same web-mentions layer. The explorer’s Online signals filter lets you narrow stories by Wikipedia, Reddit, Bluesky, Hypothesis, LibraryThing ratings, blogs, podcasts, and more.

blogssubstackEvent Dataplatform filter

June 2026 · Inferred Roles v2

Canonical role normalization

Role IDs are now normalized to a stable canonical map, reducing label drift between story pages, filters, and portfolio summaries.

roles.v2canonical IDsalias mapping

Type-aware role scoring

Role thresholds remain discipline- and type-sensitive across ArticleStory, BookStory, DataStory, and SoftwareStory. Cross-type interpretation should use evidence context, not raw score equality.

type-awareinterpretability

Expanded data/software role coverage

Added role support for data/software-specific usage patterns, including benchmark behavior, infrastructure-level reuse, and tooling/community adoption.

View DataStory and SoftwareStory role details
  • Dataset reference backbone: datasets with strong sustained citation/reuse footprints.
  • Benchmark resource: datasets repeatedly used as shared reference benchmarks.
  • Infrastructure dataset: datasets supporting broad, cross-venue ecosystem usage.
  • Software method anchor: software dependencies/methods used repeatedly in downstream research.
  • Computational reproducibility enabler: software that enables repeatable workflows and transparent methods.
  • Community tooling uptake: visible adoption in open teaching, practice, and community-facing channels.

These are heuristic role assignments and remain type-aware by design.

dataset_reference_backbonesoftware_method_anchorcommunity_tooling_uptake

Pedagogical signal expansion

Pedagogical Anchor now considers OCW, syllabi/OER, open textbook and educational video evidence across story types.

teaching signalscross-type

Repository usage surfaced for Data/Software stories

Views/downloads from repository-linked usage sources now appear in Impact metrics and Access sections for DataStory and SoftwareStory pages.

DataCiteZenodousage metrics