How we source
The method is narrow and mechanical, and it is the entire product.
Vialog exists because the same numbers circulate through this category without anyone recording where they came from. Our method is narrow and mechanical, and it is the entire product.
Every figure carries species, route and sample size
A number without a population is not a finding. Half-lives, effect sizes and thresholds differ by species, by route of administration, and by assay. Six sources publishing a half-life for the same compound can all be reporting accurately and still disagree by more than an order of magnitude, because they are describing different experiments. We record the experiment alongside the number, so a reader can see which comparison is being made before deciding what it is worth.
Claims we cannot trace are published as unsourced
When a figure appears across many sites and we cannot find a primary source for it, we list it in the claim ledger and mark it as untraced, with a note describing what we searched. We do not delete it and we do not repeat it as fact. A widely repeated number with no origin is a finding in its own right, and suppressing it would leave readers to meet it somewhere with no warning attached.
Identity data is verified, or left blank
CAS numbers, PubChem identifiers, molecular formulas, sequences and molecular weights are checked against primary chemical databases. Where we cannot confirm a value, the field reads not verified rather than carrying a figure we inherited from another site. A blank is honest; a wrong identifier propagates.
What we refuse to publish
We do not publish dosing protocols, titration schedules, cycle lengths, combinations, or conversions into syringe gradations. We do not publish vendor rankings, scores, or comparative recommendations of sellers. We do not publish testimonials, user-reported results, or before-and-after material. The first category is medical advice we are not qualified to give. The second is advertising regardless of how it is formatted. The third is anecdote presented with the authority of data.
These refusals cost us traffic. Dosing pages and vendor rankings are the highest-demand content in this category by a wide margin. We think publishing them would make the rest of the site worth less.
Corrections are public
Pages are not silently edited. Substantive corrections are logged in the corrections log with a date and a description of what changed.