Building digital-native species: Transitioning from PDF liberation to semantic-by-design publishing

What this was about

Laurence presented the 'digital-native species' workflow, noting that it takes on average 21 years from specimen collection to publication and that current practice erases structured data into PDFs only to re-extract it later. In the new model, bulk samples become a biodiversity knowledge graph linking specimens with barcodes, images and genomic data, exported as JSON to Zenodo and carried through an XML-based, semantically enhanced manuscript and peer review to multi-format publication with bidirectional links. She also showed dashboards monitoring institutions and journals, arguing that measuring a journal's FAIRness matters more than impact factor, given that taxonomic references average around 50 years old.

Why it matters. It proposes a way to shorten the long description lag and to produce certified data that can reduce AI hallucination.

In the room

• On average it takes 21 years from collecting a specimen to publishing about it.
• The old model erases data into PDFs, which then must be rescued with conversion errors.
• Digital-native species link physical specimens to digital specimens, DNA barcodes, images and genomic data from the outset, exported as JSON to Zenodo.
• An XML-based manuscript retains semantic links and identifiers through editing and peer review.
• XML-first yields FAIR-born, trusted data at the day of publication with bidirectional links.
• Dashboards (institution, GBIF hosted portal for journals, TreatmentBank) monitor impact and journal FAIRness.
• The average age of references in taxonomy is about 50 years, making 2- or 5-year impact factors irrelevant.

Notable moments