“Guadalupe”

Here’s a story of how I tried to remove a fake story marginally related to COVID-19 from Wikipedia, and, at least for now, achieved the opposite and contributed to its dissemination and perpetuation.

On a BBC-produced podcast (in Russian) I heard a story about Lupe Hernández, a nurse who allegedly invented hand sanitizer. The story was born in a 2012 Guardian article, which was subsequently quoted by viral Facebook posts and a bunch of news sites in Spanish and a bunch of other languages, and even mention in an academic nursing book published by Springer. In the last few months hand sanitizer became more popular than ever, and so the story regained popularity.

When contacted for confirmation, the original Guardian story’s author said that “she couldn’t remember the source, and that her notebooks are in storage facility she currently can’t get to”.

The podcast, as well as a thorough LA Times article, conclude that the whole story is probably an urban legend and that the person probably never existed. No one was even sure whether it’s a woman or a man, even though the original story said “she”.

The podcast did mention that there is a very short Wikipedia article. I proposed it for deletion. The result of the deletion discussion was that the article was kept and renamed to “Lupe Hernández hand sanitizer legend”.

Before it was renamed in the English Wikipedia to be an article about a legend, it was also translated to Spanish and French, as an article about “Guadalupe Hernández”, a female nurse who invented hand sanitizer, even though zero sources say that her name was actually “Guadalupe”. Sure, you can assume that “Lupe” is short for “Guadalupe”, as some imaginative writers did, but why do we do it on Wikimedia sites?

I’m still of the firm opinion that the subject should be completely removed from Wikipedia in all languages, as well as from Wikidata, but there’s only so much I can do about this. If any of you know French or Spanish, can you please make sure the articles in your languages are not too awful, or perhaps consider proposing them for deletion?

And if you think I’m badly wrong about it all, please do tell me, too.

The Persistence of Decision Points in the History of Wikipedia

The Persistence of Poverty, which is today’s episode of NPR’s famous “Indicator” podcast, made me think of how small things that happened long ago in the history of Wikipedia and other Wikimedia wiki sites still affect us, for better or worse.

Here are some examples.

Example one: People didn’t want to have full copies of historical documents on the English Wikipedia, because they are not encyclopedic articles. So they created a whole separate wiki for it, called “Primary Sources Wikipedia”: “ps.wikipedia.org”. It turned out that this would be the URL for the Wikipedia in the Pashto language, which has the ISO 639 code “ps”, so it was renamed to Wikisource, becoming Wikimedia’s first non-Wikipedia wiki. The movement wasn’t even called “Wikimedia” then—the organization was created later. Later, Wiktionary, Wikibooks, WikiCommons, and other projects joined. And Wikisource and all of these other projects are awesome, but now this also has the side effect of having to have some challenging discussions between the Foundation and the community about how non-Wikipedia wikis should be branded in the long term.

Example two: A French Wikipedia editor who is curious about Ancient Egypt wanted to insert Egyptian hieroglyphics into Wikipedia articles, and he happened to know some PHP, so he wrote the Wikihiero extension, which is installed on all the wikis. Because it’s an extension that adds its own wiki syntax, Visual Editor shows a button to insert Hieroglyphics on every page, including the page about Astronomy on Wikiversity, which doesn’t have much to do with Ancient Egypt. This is not bad—this is mostly very good. What is bad is that the Visual Editor doesn’t have a button to insert infoboxes or “citation needed” tags, even though they are far more common than hieroglyphics, because they are implemented as templates and not as PHP, and Visual Editor handles all templates as one generic type of object. (If you are wondering how can this get fixed, the first necessary step in that direction is described on the page Global templates on mediawiki.org.)

Example three: Some people didn’t like that too many wikis are created in new languages and stay inactive, so they wanted a proper way to prove that people plan to be active editors. So they created the “Incubator” wiki, where people would show they are serious by writing the first bunch of articles. For various technical reasons, using it was more difficult than using a usual Wikipedia, but they probably quietly assumed that everybody who wants to create a Wikipedia in a new language is experienced in editing Wikipedia in English or Italian or some other big language, so almost no one ever bothered to improve it. By now, we know that that assumption was tragically wrong: most people who want to create a Wikipedia in a new language are not experienced in editing in other languages, so they are newest and the least experienced editors, but they get the most complicated user interface. (If you are wondering how can this get fixed, see this page on Phabricator.)

Yes, I’m oversimplifying all of these stories for brevity. And I’m not implying any malice or negligence in any of the cases here. These were good people with good intentions, who made assumptions that were reasonable for the time.

It’s just a shame that the problems they created are proving more difficult to fix as the time goes by.