Methods and data quality

Collection

Each entry starts from a name on an editorial list. For that name the site retrieves the compound record from PubChem, a literature search from PubMed and the article summary from Wikipedia. Where a name alone finds the wrong record, the entry is tied to the identifier of the record instead.

Chemical identity

Formula, molecular weight, other names and the structure drawing are shown as PubChem records them. A CAS number is accepted only when its check digit is valid. Registry codes from other databases are left out of the list of names, and a name known to be wrong can be excluded by hand.

Published literature

By default the PubMed search looks for the name in titles and abstracts; a few entries use a wider search, and every reference sheet prints the query it used. A short selection of the first results by relevance is listed with authors, journal, year and type of study, next to the total number of results. Titles are not translated or edited.

Summaries

The presentation is one sentence, the first of the Wikipedia article, with a link to the article and its licence. In other languages the sentence comes from the Wikipedia in that language when it has the article; otherwise the English sentence is shown and marked.

Dates and updates

Every reference sheet states the day its data was retrieved. The data is refreshed by running the collection again, and nothing is edited by hand in between. A database can be out of date or wrong: check the original record before relying on a value.

What is left out

This site writes no statements about effects, benefits or safety, and gives no information on how any substance is used. It names no sellers and links to none. A substance being listed says nothing about its quality or legal status.

See About for the sources.