Saturday, August 16, 2008


Wikipedia is often clubbed together with open-source software (such as Linux) or open-access publishing (such as the PLoS journals) as an example of how information wants to be free. But, as many have pointed out, there is an enormous difference. Open-source software such as Linux, and scholarly journals such as PLoS, are rigorously peer-reviewed.

You can take Linux code and do what you like with it, such as adding a device driver for your own obscure hardware, within the restrictions of its licence: the GNU General Public License. You can reproduce PLoS articles and use their content in other ways, within the restrictions of their licence: a Creative Commons licence. What you cannot do, without satisfying extremely high standards of quality, is have your own code included in the Linux sources distributed by Linus Torvalds, or have your own research published by PLoS.

With Wikipedia, however, anyone may edit an article, and it is instantly readable by millions.

Surely the difference is glaringly obvious, but defenders of Wikipedia have claimed for years that mischievous edits are quickly reverted by editors. And this is often true, but I suspect that it is now much less true than formerly. Where it used to be rare to find a Wikipedia article, on a subject of even slight general interest, that had glaring errors or obvious "ego" additions, it is now rare to find an article that does not have such defects.

A couple of years I found Wikipedia a useful and largely reliable resource. Now I find it a useful resource that must be double-checked on every point, no matter how minor.

Here is a list of recent examples:

  • Some time ago Slashdot linked to a Donald Knuth interview, where he talked of, among other things, literate programming -- and linked to the Wikipedia article on it. Unfortunately, as of that date, that article was so incorrect as to be worse than useless (and sparked an ill-informed flamewar on Slashdot too). After the Slashdot link appeared, an "edit war" ensued and it seems to have been restored to a somewhat informative state; but the incorrect version was what was available for months earlier. It used to be that one could trust Wikipedia on technical subjects, but that seems no longer to be true.

  • When I checked Wikipedia for "semolina" some months ago, the following sentence [UPDATE 18/08: link fixed] caught my eye (and it was in the references section of the article!) "Semolina in India is known as "Sooji". Various delicious preparations are made with Sooji. Sooji Halwa is the most famous which is a sweet item. Among Bengali's also this is very popular and used for preparing various sweet items. One of them in Bengali cuisine is known as Soojir Payesh, which is prepared by boiling sooji balls in sweetened milk." Not incorrect, but hardly very professional-looking, and not in the appropriate section; but this text survived from March 25 until May 12.

  • Some time ago Rahul Basu observed, with amusement, that there was a Wikipedia entry for the name Rahul; and regretted that he did not qualify for the list of famous Rahuls. What he did not notice was that several others had had the same regret, and attempted to amend matters; the edit history of that page is rather interesting, and the current choice of "famous" Rahuls is quite strange. (Any such list seems rather pointless.)

  • And finally, my most recent peeve and the one that provoked this post: in the article on the Wodehouse character R. Psmith, Wikipedia (as of today) claims that his name, in the last book (Leave it to Psmith), was changed from Rupert to Ronald Eustace. The "Ronald Eustace" appellation dates from Wikipedia's very first entry on the subject, in March 2005. Now, the second question to ask is, "why did Wodehouse do this?" and Wikipedia suggests that it was to avoid confusion with another Rupert in the book (Rupert Baxter).
    But the first question to ask is, "is this true?" Leafing through my copy of the book, the only evidence I can see to back this claim is the following exchange towards the end of the book, between Eve (who had been taken in by Psmith's impersonation of the Canadian poet Ralston McTodd) and Psmith:
    "Mr..." She stopped. "I can't call you Mr McTodd. Will you please tell me your name?"
    "Ronald," said Psmith. "Ronald Eustace."
    "I suppose you have a surname?" snapped Eve. "Or an alias?
    "There's not much sense in pretending now, is there? What is your name?"
    "Psmith. The p is silent."
    Personally I cannot see how this could be read as indicating that Psmith's given names were "Ronald Eustace": he was obviously "pretending". But this bit of folk wisdom seems to have become replicated all over the internet, never to be stamped out.

Colbert was right: soon our reality will be defined by Wikipedia.


km said...

I have pretty much stopped reading the wiki entry for most of my googled terms.

//Though, as someone will undoubtedly remind me, if one doesn't like a Wiki entry on a topic, one should do something about it.

Rahul Siddharthan said...

km - if you "do" something about it, your changes will probably be reverted, unless it's either a completely purposeless (or, as they say, "neutral") change or it's a point of fact for which you cite a very impressive and easily accessible source. On the other hand, if you make a clever mischievous edit, it may well stay undetected forever. The subject of a later post, perhaps.

Rahul Basu said...

Rahul: your link to sentence in the note about Sooji takes one back to the top of your blog page rather than (presumably) to the entry in Wikipedia. Incidentally, apart from a totally gratuitous apostrophe in "Bengali's" , as far as I know, the description of Soojir Payesh is not even correct - there are no Sooji balls, only Sooji in sweetened milk (unless the dessert has changed since I had it last!).

BTW the Semolina/Sooji entry no longer has all this pointless stuff!

Rahul Siddharthan said...

Rahul: link fixed, thanks. I plead ignorance about the details of soojir payesh :)

Akilan said...

>>if you "do" something about it, your changes will probably be reverted

Yes it is. But the problem is as the sysops and good editors are reverting the nonsense, there are others who write new nonsense at a higher rate. Some pages are more susceptible to this kind of repeated vandalism. Eg. biographies of living people, companies etc. Kalam's wiki page is the classic example.