We have not written a standard

The word has an owner

In ordinary speech, "standard" means something that is widely used and generally agreed. In the world we operate in, it means something narrower and much more specific.

A standard is a document published by a recognized standards body after a formal process. That process has defined stages: a committee with declared membership, national delegations, successive drafts, a public comment period, and a ballot. It takes years by design.

Its authority does not come from being correct. It comes from how it was made — from a procedure that anyone can inspect, that gave interested parties a route to object, and that recorded what happened when they did.

A specification is a document that says precisely how something should be done. Anyone may write one. Its authority comes only from whether people find it useful enough to adopt.

Common Measure has written a document. It has not run a balloted international process with national member bodies. So what we have is a specification, and we should say so.

Standards are not better documents. They are documents that went through something.

Three reasons we are careful about this

The first is simply that we would be caught.

Say "we have published a standard" in a room containing anyone who sits on a technical committee, and the questions arrive immediately. Which body. Which committee. What was the ballot result. Who were the national delegations.

There are no good answers to those questions, because we have not been through that process. The claim would fail in the one room where credibility matters most, and it would fail in the first thirty seconds.

The second is that ISO is already here.

ISO 17805:2026 specifies how environmental DNA is sampled, captured and preserved from water. It was published in 2026 and adopted in Europe through CEN. Further standards covering extraction from filters and reporting requirements are in development, and in Canada, CSA W214:21, Environmental DNA (eDNA) reporting requirements and terminology, is a published National Standard of Canada.

If we call our document a standard, we are implicitly claiming to be in the same category. A reasonable listener hears us positioning against ISO. That is a fight we could not win, should not pick, and do not want — we adopt ISO 17805 by reference rather than restating it, and we would rather contribute to that work than run alongside it.

The third is that the honest label is more useful.

"Specification" tells a reader exactly what they are holding: a document written by an identifiable party, published openly, not yet blessed by any process, to be judged on its merits. That is an accurate description and it sets the right expectation. Nobody who adopts it can later feel misled.

What this frees us to do

Being a specification is not a lesser version of being a standard. It is a different instrument with different properties, and several of them are advantages.

We can publish in weeks rather than years. Version 0.1 of the assurance record went out a month after it was drafted. A balloted process could not have done that, and should not try to.

We can admit what we have not settled. The specification carries a section of open questions — whether a field should be mandatory, how multi-marker runs should express the marker, whether conformance should have levels. A ratified standard cannot say "we do not know yet." Ours can, and the honesty is worth more than the polish.

We can be corrected quickly. When someone shows us a field is wrong or impossible to populate in a real workflow, we can change it, publish the change, and record why in the change record. No ballot required.

We can ship working code. A machine-readable schema, a validator that runs in a browser, a crosswalk to Darwin Core and MIxS. Standards bodies publish documents; that is their job. Somebody also has to build the tools.

And we can be openly licensed. Everything we publish is free permanently under CC BY 4.0 — free to read, to check, to implement, and to build commercial products that compete with us. Many formal standards are paid documents, for legitimate reasons relating to how standards bodies fund themselves. We are not constrained the same way.

So what is Common Measure, precisely

Three things, and it is worth separating them because they are often collapsed together.

A specification for a narrow gap. Not for environmental DNA as a whole. Sampling and preservation are governed by ISO. Laboratory competence is governed by ISO/IEC 17025. Reporting terminology is covered by CSA W214:21, Environmental DNA (eDNA) reporting requirements and terminology, a published National Standard of Canada. What no published standard currently addresses is what happens to data after it leaves the sequencer: how a sequence keeps a stable identity across runs, laboratories and years, how reference database versions are tracked, and how far apart two laboratories actually are. That is what we specify, and only that.

An operator of the infrastructure that applies it. A specification is a document. Somebody has to run the thing that takes real data, at volume, and produces results in a form other systems can consume. Writing and operating are different jobs and we think they should sit with different organizations — which is why we do not run samples and never will.

A governance structure that keeps the first two honest. The decisions inside the specification — which reference databases, at which versions, how names map to identifiers when taxonomy is revised — are placed with an external science committee on which we hold no vote. Not because it sounds good, but because a company that both sets the rules and sells the service has an unresolvable conflict, and everyone can see it.

Where this is supposed to end up

The sensible long-term home for this work is a formal standards body running a proper process.

We are not being modest. If ISO, CSA or an equivalent takes up sequence identity, reference version control and cross-laboratory comparability, and publishes a standard covering them, then the specification we maintain becomes unnecessary and we should retire it. We would regard that as the system working, not as being displaced.

What we would still do afterwards is the operating part — running infrastructure that implements whatever the standard says, which is the ordinary relationship between a standard and the organizations that apply it. Accounting standards have auditors. Measurement standards have calibration laboratories. Nobody thinks the auditor and the standards board should be the same organization.

In the meantime, standards processes take years and data is being produced now that will not be comparable to anything. Somebody has to run something in the interim, publish what they learn, and hand the evidence to the bodies doing the formal work. We think that is a more useful position than waiting, and a more honest one than pretending to be something we are not.

If you take one thing from this

When you read that an organization has published a standard for something, it is worth asking which body and what process.

Sometimes the answer is a real committee, a real ballot, and a reference number you can look up. Sometimes it is a company that wrote a document and liked how the word sounded.

Both can be useful. They are not the same thing, and the difference is worth knowing before you build anything on top of either.