Gemacode Research works on one question, from several directions: how can a record stay verifiable after the systems, vendors and organizations that produced it have changed?
Storage is solved. What is not solved is proving, to somebody who has no reason to trust you, that what you show today is what existed then — and on what evidence a conclusion was reached. Every public line below is an attempt on that problem in a different domain.
Each line publishes something a third party can inspect and reproduce — a protocol, a graph, a standard — rather than a claim about our competence.
How a declaration can be fixed so that its integrity and authorship survive the organization that issued it — and how heterogeneous sources can be verified and weighted before they count.
How to trace, from a declared composition, which biological pathways are documented — attaching an evidence level and a source to every connection, and recording documented absence as absence.
How a sector standard can report what is documented about a product without becoming a safety verdict — and how each assessment becomes a record an outside party can verify.
Four commitments that apply to every line — and that we would rather be judged against than believed on.
A specification, a conformance suite, a signed record, a source-tracked dataset. If a claim cannot be handed to someone as an inspectable object, it is not finished.
Every system here documents what it does not prove, in the same place it says what it does. A boundary written down by us is worth more than one discovered by a reviewer.
No identified evidence means exactly that — never approval, never safety. This single rule shapes both the biological work and the evidence protocol.
Conclusions weigh corroboration by genuinely independent sources rather than the standing of whoever asserts them. Trust is computed and re-checkable, not granted.
The test of this method is whether an outsider can pick up the work without us. These artifacts are public, free to fetch and reproducible.
The same commitments stand behind every public system — that is what makes them an infrastructure rather than a set of claims.
Every public record can be checked by a third party, without relying on Gemacode or its servers.
Claims carry sources and evidence levels. Absence of identified evidence is reported as absence — never as approval.
Open protocols and an open verification model, so the ability to verify outlives any single vendor.
This is the public scope of Gemacode Research, not an inventory. Some research lines are deliberately not described here — including work that is not ready to be judged, and work whose publication is not in our interest to rush.
The most useful thing you can do is try to break something. Implement the protocol from the specification and report where the text is ambiguous; run the conformance vectors against your own code; challenge a published record; tell us where a stated boundary is wider than we admit.