Greyscale Gulls
In researchLearning to identify gulls
A course and an identifier for the hardest common birds in Europe. The teaching unit is not the species but the plumage stage — species by age by season. That is exactly why photo-to-species apps break on gulls.
1What it is
To write
What it is: a course, and an identifier used as its exam.
2Why the fast answer fails
To write
The fast answer is photo → species, guessed. Write out why it breaks here: a four-year gull is eight or nine distinct looks, so two birds of one species share nothing and two species look like twins.
3The hard part
To write
The taxonomy, not the model. The atom is species × age × season, not species — 29 Western Palearctic species, each with its cycles and seasons, and every exemplar photograph needing a clean licence and a confirmed identification.
4Product decisions
To write
Three to five decisions with their trade-offs. Candidates: the plumage stage as the unit, which multiplies the content work; teaching age before species; a geography and season prior that narrows the answer set; front-loading the common species by all their stages rather than covering many species shallowly.
5Where it is now
To write
Honest status: species list, schema, mantle-grey method and a course draft exist. The app does not. The blocker is verified reference photographs, not code.
6Screens and artifacts
To write
The species table, the plumage-stage schema, a sample lesson.