Where an open version would matter
The registry records what exists. This is the other half: places where the closed version is the only version, the reason it stays that way, and what an open alternative would actually have to produce.
This is editorial, not a record. Registry entries are cited facts about things that exist; an opportunity is our argument about something that does not. The links to existing work are checked against the catalog — the judgment around them is ours, and you should disagree with it where you know better.
Measured from the catalog, not asserted. A thin category usually means the work is happening somewhere our importers cannot see — open hardware and clinical guidelines rarely live in a Git repository — rather than that nobody is doing it.
Open designs exist in quantity. Almost none can be legally manufactured and sold, because the barrier is a documented quality system, not a CAD file.
hardest part: regulatory · manufacturing · capital
Treatment is lifelong, three times weekly, and consumes a large share of some health systems' budgets. Machine and consumable design is closed and the recurring cost is structural.
hardest part: engineering · regulatory · manufacturing · reimbursement
Medical oxygen is a basic input that fails at the maintenance layer. Devices arrive as donations, break, and cannot be serviced because parts and documentation are closed.
hardest part: engineering · manufacturing · distribution
Locally made devices are useless if they cannot be sterilised. Autoclave availability and validation, not device design, is often the real constraint on surgical capacity.
hardest part: engineering · regulatory · distribution
A diagnostic test is mostly reagent cost, and reagents are imported, cold-chained, and priced far above production cost. Local enzyme production is a solved science with no distribution.
hardest part: manufacturing · distribution · capital — 1 related entry
The signal processing is well understood and publishable, the hardware is commodity, and regulatory access in the US recently opened. Almost nothing open exists.
hardest part: engineering · regulatory · distribution — 1 related entry
A century-old biologic with a concentrated supply chain. The molecule is not the obstacle; the manufacturing and regulatory route for a small producer is.
hardest part: manufacturing · regulatory · capital · science — 1 related entry
Isothermal amplification removes the need for a thermal cycler and is widely published. Commercial point-of-care systems still bind each test to a single-vendor cartridge.
hardest part: science · manufacturing · regulatory — 2 related entries
Passive 3D-printed hands are a mature open success. Powered, sensor-driven limbs remain closed and cost orders of magnitude more, and the control software is the gap.
hardest part: engineering · science · distribution — 1 related entry
A treatment planning system is a large per-site licence bound to one manufacturer's linear accelerator, and the dose physics underneath it is textbook material.
hardest part: regulatory · engineering · capital — 3 related entries
The open diabetes community solved the software and left the sensor closed. The recurring cost of consumables, not the algorithm, is what keeps continuous monitoring out of reach.
hardest part: science · manufacturing · regulatory — 5 related entries
Most of the world has no access to MRI. Low-field designs are physically simpler and already exist as open prototypes, but nobody has carried one through manufacture and clearance.
hardest part: manufacturing · regulatory · capital — 7 related entries
The most useful response to this page is a correction. If a gap here is already solved and we missed the project, that is a registry bug and worth an issue. If an entry is here because it sounds compelling rather than because it is tractable, say so — an opportunity list that nobody argues with is decoration.