Design an Organism Knowledge System
Collect observations offline, build a reviewed organism catalog on Earth, and prepare guidance for the next human expedition.
Problem statement
You are the first explorer to visit a planet with life. Design an information system for three stages: record the organisms you encounter, turn the returned observations into a more complete biological information system on Earth, and help the next human crew learn from it.
For this exercise, assume one explorer uses one device with no network during the trip. Each observation has a stable ID, a time, a rough location, notes and optional photos. A successful local save means its complete payload is durable and survives an app restart.
If storage is full, the app must not claim that the save succeeded.
For this exercise, assume Earth imports can be interrupted and retried. Importing the same observation ID and payload again must not create another observation. Reusing an ID with different content must be detected rather than silently replacing evidence.
Pending metadata may be shown for import status and progress, but only complete verified imports may enter researcher curation. Researchers keep raw observations unchanged, review organism interpretations with links to their evidence, and publish immutable catalog revisions.
For this exercise, assume the next human crew downloads readable guidance and selected photos from one chosen published catalog revision for offline use. A pack must not mix revisions or silently change after publication; a correction produces a new catalog revision and a new pack.
Cover app restarts and interrupted transfers. Loss of the entire device, autonomous rover control, automatic species classification and a general course platform are outside scope.
What to build
- Capture and read complete field observations offline, preserving every acknowledged save across app restarts.
- Import returned observations with visible progress, safe retries and a clear distinction between partial transfers and complete evidence. Pending metadata may show transfer progress; only complete verified imports may enter researcher curation.
- Let Earth researchers revise organism interpretations and publish a catalog without losing the unchanged field observations and their provenance.
- Prepare and download human-readable guidance with selected photos pinned to one published catalog revision for offline use.
At scale
- For this exercise, assume at most 20000 observations per expedition and at most 2 MB of total text/photo payload per observation. Plan local capacity for the bounded payload plus storage overhead.
- For this exercise, assume no network during the trip. Once an observation payload is ready, target a successful durable local save within one second; an app restart may occur after any acknowledgement.
- For this exercise, assume Earth stores 1000000 observations and supports 100 researchers reviewing or editing at the same time.
- For this exercise, assume a return transfer may stop after any request, and a published pack must remain on one immutable catalog revision even after later corrections.