Wikipedians not amused

06 February 2017 - Wikipedia in the literature

wikipediachemicalstructureexplorer.PNGThe chemistry Wikipedians have had a difficult week! Someone pointed them in the direction of a paper that, although already published in 2016 took them by surprise. Titled "Glaring Chemical Errors Persist for Years on Wikipedia" in the Journal of Chemical Education and authored by Michel Mandler (DOI) it basically states that the Wikipedia chemistry section is rubbish in chemical structure drawings. In the ensuing discussion here it is mentioned that the errors found are "pretty obscure examples" and correcting errors "often took hours". The most damning comment is that the author "looked for a shortcut to his 15 minutes of fame by belittling a community he knows nothing about". Other commentators let it know they did not have a beef with the author.
In the discussion there was another mention of an article about Wikipedia (this one from 2015 and certainly unknown to me) and this one was definitely favourable! In it Peter Ertl et al. describe a chemical substructure search engine based on the Wikipedia chembox templates (DOI). The thing is, Wikipedia is not just an encyclopaedia, it is a database of molecules as well and easily extractable. The practical result is search engine page at http://www.cheminfo.org/wikipedia/ preloaded with a 2 MB json file containing the entire Wikipedia chemical structure collection (15200 entries). Combined with a Javascript molecular editor and substructure search software also based on JS it is a complete frontend search engine. Examples: a search for brome-hexane also yields bromoadamantane and BOMT. The oxetane group also includes diketene, malonic anhydride and merrilactone A. The article describing the search engine does report the Wikipedia chemboxes contain errors but apparently to a manageable extent.