A short controlled lexicon. The words n1lab uses, the words it avoids, and the difference between them — annotated.
Health software is loud. It uses urgent verbs, asserts conclusions, and rewards behaviour. n1lab does none of that. The vocabulary below is the working list — every word in the product, every entry in a tooltip, every line of copy is checked against it.
If a word you'd expect to see is missing, that is usually deliberate. If it appears here marked avoid, the reason is in the entry.
A repeated co-occurrence between two recorded variables. Used in place of cause, which the data cannot support.
A trial closed before all three windows were logged. Never failed; the word "failed" implies the trial owed you a result.
The fourteen-day rolling mean of a given variable. Always rendered as a shaded band — mean ± 1σ — never as a single line. The baseline is yours and it moves.
The 0–10 score for a symptom domain at a given reading. Used in place of severity, which carries a clinical weight the score does not earn.
Belongs to a vocabulary that treats the body as something to be defeated. n1lab does not.
A three-level scale — low, moderate, high — paired with an explicit agreement count. Never used without the count. "High confidence" alone is meaningless; "high confidence · 4 of 4 observations agree" is a sentence.
A variable recorded at the start of a trial that may influence the outcome — poor sleep, alcohol, a new medication. Confounders are recorded honestly; they are not held against the trial.
A reading above the upper edge of the baseline band. Used in place of abnormal or bad, which assert a judgment the data does not carry.
The set of readings that support a recorded observation. Always quantified — three of four, two of four — never adjective alone.
Asserts a teleology the data cannot see. Use "symptoms decreased" or "burden returned to baseline."
A trial with mixed signal, confounded readings, or insufficient observations. A valid outcome, not a failure.
A clinical term n1lab does not have the evidence to use. Replace with "reaction observed."
A single recorded reading. The most-used noun in the system. Plural. Quantifiable. Suggestive on its own; meaningful in series.
Implies a known maximum. There is no known maximum for any variable n1lab tracks. The body is not a parameter.
The default modifier on any association the app surfaces. Drops only when at least two trials have agreed.
Symptoms increased above baseline within the 24-hour window. The canonical phrasing for a positive trial. Never shortened to "reaction" alone; the verb "observed" is what makes the phrase honest.
A clinical assertion the app does not make. Replace with "tolerated."
A pattern visible above the baseline noise. The thing repeat trials strengthen.
n1lab does not reward consistent recording. Streaks would convert the instrument into a game. There are no streak counters anywhere in the app.
The strongest modifier the app applies to a single trial. Two trials agreeing promote a finding from suggestive to working — never to conclusive.
Symptom load stayed within the 14-day baseline across the trial day. Never safe; tolerated says only that this trial, with these confounders, in this body, did not produce a recorded reaction.
Belongs in a different software category. n1lab is recording, not metamorphosis.
One of the three intra-day intervals on a 24h food trial: 0–2h, 2–6h, 6–24h. The interval inside which an observation is made, not the boundary itself.
A category of software that overlaps n1lab's surface but not its method. Excluded throughout the product, the marketing, and the documentation.