Skip to main content

2.5.1

· 3 min read
André Aubin
André Aubin
Trefle Core Team

Two fixes for defects that shipped with 2.5.0, one of which was doing exactly the opposite of what it was for.

You can read the terms you are asked to accept​

2.5.0 asks signed-in visitors to accept the current terms of use before continuing on the website. The acceptance form offers a link to read them first — and that link bounced straight back to the form.

The gate exempted the acceptance page but not the terms themselves, which are served by a different controller. So the one route offered for reading the document was the one route that could not be reached while the gate was up. Asking someone to agree to a document they cannot open is not a defensible thing to ship, and it is fixed.

API access was never affected: the gate has always ignored /api/v1 entirely, and tokens kept working throughout.

WFO now declares its licence​

The licence work in 2.5.0 seeded three of the four sources it knew about. The fourth, World Flora Online, is stored internally under a differently-cased identifier than the one the seeding step looked for, so it was skipped — silently, and reported as a success.

WFO is one of the largest taxonomic sources behind Trefle, so a good share of sources[] entries were carrying licence: null when the licence is in fact CC0-1.0. It is now filled in:

{
  "name": "WFO",
  "licence": "CC0-1.0",
  "licence_url": "https://creativecommons.org/publicdomain/zero/1.0/"
}

The seeding step now also reports which sources it could not find rather than passing over them, so a run that half-succeeds stops looking like one that worked.

Naming the authority​

One documentation change worth mentioning, because it affects what you can expect from the data. Trefle now states plainly which source decides what a plant is called: the World Checklist of Vascular Plants (Kew) arbitrates accepted names and synonymy. POWO is its web front end and may lag it; IPNI records that a name was published but never that it is accepted; GBIF aggregates. None of those overrules WCVP.

The same document records the nomenclature we store — subsp. rather than ssp., × rather than the letter x — and one distinction that is easy to get wrong: an autonym such as Stipa pulcherrima subsp. pulcherrima is a narrower taxon than Stipa pulcherrima, not another way of writing it. Data may be generalised from the autonym up to the species; it must not be pushed the other way.