Wikipedia needs flagged revisions, so that anonymous edits are approved by someone who is actually committed to the idea of building an encyclopedia, rather than to malice or lulz.

Here is an example: half the internet (and at least one book on haircare) thinks that Erica Feldman or Ian Gutgold invented the hair straightener, based on schoolkid vandalism in Wikipedia 6 years ago.

It's listed in the list of Wikipedia hoaxes and is also one of the top 5 trolls listed on "House of Geekery".



It is back in Wikipedia right now:



Of course, it cites a source: a website that copied it from Wikipedia. (The actual inventor of the hair straightener, before the Wikipedia vandalism, was Madam C J Walker.)

Wikipedia is spreading lies as well as knowledge. With 4 million articles, editors are stretched much too thinly to ensure quality control under the present set-up.

Also see previous discussion on Wikipedia Review.

So you propose fact-checking by experts using their own name and credentials? Are you serious?

I used to publicise errors like this, but then people would make the corrections and claim that the correction mechanism works after all, in a convoluted way. I simply keep notes now. There is one article, in my area of specialism, that is being increasingly vandalised and I am interested when anyone is going to pick it up. This is not vandalism that needs an expert, either. A simple consistency check would show that at least one part of the article is wrong.

No, just editors with a track record of bona fide work, who wouldn't approve edits like the Feldman or Gutgold edits.

Real-name registration would be one step further. Nice too, but one step further. But flagged revisions work without that in the German, Polish and other Wikipedias.

On the Swedish wikipedia, I found a hoaxer who had really put a lot of effort into editing. He would have become a trusted user, whose edits would not need approval.

http://sv.wikipedia.org/wiki/Diskussion:Itm%C4%81m by user Josephus (later Sofokles, I do not know under what name he operates now). He claimed to know Arabic, but when a reader protested that there were blasphemous phrases in the Swedish articles about Islam, he denied understanding what he had written. And of course the community did not really throw this guy out.

He had also edited on enwp, to create support for his Itmam hoax, things like http://en.wikipedia.org/w/index.php?tit ... =191251088

Nice. Reminds me of the Hannibal Fogg hoax. Someone like that would no longer be a trusted user, and never again become one.

May I put articles about these on the wiki?

I do not quite understand the question. If you want to have something about the Swedish Itmam hoax, I can contribute.

On enwp, there is SergeWoodzing, inventing "English" names for Swedish royalty and historical stuff.

There is a lot of Frisian nonsense Gosse Ludigman never existed, see also the rest in http://en.wikipedia.org/wiki/Category:P ... _Friesland

Even more glaring is that even well-known English history can be so wrong. The natives had not spotted that "Honour of Richmond" with strange quotes attributed to William the Conqueror was a complete fake in 2007. It is now deleted, but the talk page was restored http://en.wikipedia.org/wiki/Talk:Honour_of_Richmond

So who would trust articles about medieval Balkan states, or kingdoms in India?

Please do!

I am personally aware of at least five detailed articles on non-existent towns and villages around the world. Documenting imaginary towns is pretty benign, as disinformation goes, but proves the point of Wikipedia's underlying unreliability.

As an aside, the end of most hoaxes is when they are revealed by the hoaxer in order to achieve credit or to make a point. The flip side of this is that we have no idea how many deep-cover or forgotten hoax articles are lurking in Wikipedia at any moment.

I'd do this a lot...if I gave a fuck.

I inserted a hoax into Wikipedia, to support a presentation I later delivered to Cheltenham High School. There was no radio station planned for the school, and over 2,200 page views of the article over nearly 7 weeks didn't prompt a correction. I am convinced that had I not revealed the hoax to the school on the day of my presentation (which prompted reversion that evening, presumably by one of the students), the hoax would still be in place now.

"Beginning April 1st"...

I share your view. I am likewise sure that if I had not mentioned the Feldman/Gutgold thing the other day on the CREWE Facebook page and on the en:WP mailing list, it would still be in there, and be in there for the next five years. Typically, when I find a thing like that, it is clear no one else cares. There are literally millions of Wikipedia articles that are vulnerable to this sort of thing, every day.

I myself know a friend who invented a fictitious 19th century economist earlier this year. The article still stands complete with fake/misleading sources. In fact he even told me that, if he wanted to do so, "I actually have a handful of daguerreotypes of random people. I could add [a picture to the article], easily."

You know what else has happened to his article? The same phenomenon the WikiWatch guy once noted happened to a fake article he was monitoring in 2005:

I believe I've found the article you speak of.....will PM you.

No, that isn't the person. I've sent you a link to my friend's article.

Don't forget the "official" list of hoaxes. Pathetically inadequate, always out of date.
Note that Gaius Flavius Antoninus was there from 2004 until last month. Didn't anyone notice it before? It's the oldest one on that list
but hardly the oldest one on Wikipedia generally.

I'll still take Semi-protection any day over Flagged Revisions.

*process goes like this with flagged revisions - IP makes and edit (usually unsourced) and disappears - sometime later reviewer comes along and sees edit, has no idea whether it is ''bona fide'' or not unless they hunt for sources themselves.....and can't ask IP who has long gone.

*Now with semi-protection, IP makes request, admin or established editor can discuss veracity/request sources etc. and IP is (hopefully) there and/or keen to provide rationale. Hey presto, someone provides a source

No brainer really......

So, you're saying I'll have to wait 4 days and make 10 edits to unprotected articles before I'll be allowed to execute a masterful campaign of fraudulent edits? You set the bar so high, Cas!

If flagged revisions passes, there is really nothing to stop the admins giving the ability to pass revisions to absolutely everyone. The dwindling admin corps are not about to police every edit.....

At that point it will become meaningless. Just another (worthless) layer of hat-collecting. While the real damage will still be prevented by semi & full protection.

Why didn't this demise occur on the German Wikipedia?

If I remember correctly, anyone with over 300 revisions automatically gains the right to review revisions on dewiki.

Which part are you referring to? I was also under the impression on DE it was auto-assigned after a set amount of revisions, if you mean 'why hasnt it become meaningless on DE' well it would be better to ask someone on DE. But DE isnt EN. The Germans are not quite so batshit insane.

Passive (autoreviewer) and active reviewer rights are indeed auto-assigned, but they can also be withdrawn at any time.

Welcome.

I'd say that a large number of people who make unsourced edits wouldn't have a source. In the cases where they do (because they just read it in the paper), it's trivial (and faster) to Google the fact in Google News than to leave the IP a message asking for a source.

The advantage of flagged revisions is that the feature covers (or could cover) all articles (or all BLPs), and prevents immediate display of unsourced and unchecked content to the public. Semi-protection is quite rare.

People can (and do) drive trucks through Wikipedia's present quality control set-up.

Yeah, but if said account starts misbehaving they can be blocked......

Also WRT semi-protection being rare, I've always used it pretty liberally since 2007 and been happy to engage in discussion on the talk page and unprotect and watch closely....can't comment with authority on JN's assertion that most additions are unable to be sourced (it is true that google searches are often frustratingly fruitless I must admit......)