The Angry Corrie 65: Jul-Sep 2005

Grant Hutchison is ... The Wiki Man

Wikipedia is a rather odd notion the first time you encounter it: an "Open Content" online encyclopaedia, maintained by a very large number of enthusiasts who type up entries on anything that takes their fancy. After that, the Open Content idea becomes important - it means that anyone who notices an error or omission in a Wikipedia entry is free (indeed, encouraged) to edit the entry to make it better. So the ultimate extent and accuracy of the encyclopaedia are emergent properties of the endless tweakings of thousands of knowledgeable enthusiasts. Wikipedia is a specific example of the general concept of wiki - web resources that allow anyone to add or edit content. (In case you're wondering, wiki comes from the Hawaiian phrase wiki wiki, meaning "quick" - a slightly affected little coining, I can't help but feel, given that the coiner, Ward Cunningham, isn't actually Hawaiian.)

But what stops Wikipedia being filled up with rubbish by idiots and vandals? Nothing at all, say Wikipedians, except that there are more good and sensible folk out there fixing bad content than there are idiots and vandals putting it in. "Given enough eyeballs," they claim, "all errors are shallow." (In the spirit of wiki, this is a revised rendering of something called Linus's Law, which referred to the mass debugging of the Linux operating system: "Given enough eyeballs, all bugs are shallow.")

But: "Given enough programmers," I hear you mutter, "all philosophy is shallow." So I'll crack on with the business - does Wikipedia deliver what it promises? It's been around for five years now, and contains half a million articles, so the wiki process should have had time to come up with some good solid information about mountains.

To give it a fair outing, I thought I'd draw up my terms in advance. First, I made a varied list of 20 mountains. In no particular order, these were: Aconcagua, Elbrus, Kilimanjaro, Denali, Everest, Vinson Massif, Mount Wilhelm, Mount Kosciuszko, Mont Blanc, K2, Roraima, Kanchenjunga, Ben Nevis, Gunnbjørnsfjeld, Chances Peak, Chimborazo, Mount Cook, Yding Skovhøj, Rysy and Mount Olympus. Then I jotted down ten things I would hope to find in a good encyclopaedia entry about a mountain: height, origin of name and/or alternative names, latitude and longitude, geography, geology, history, biology, first ascent, routes of ascent, and a picture.

Only then did I visit Wikipedia and start typing names into its search engine - since Wikipedia is a distinctly dynamic entity, it's important to record that this happened on the evening of 8 May 2005. I went straight for the core entry for each mountain, and didn't follow any internal or external links offered. Then I sat down with the 2002 paper edition of the Encyclopaedia Britannica, and looked up the same 20 mountains; again, I went straight to the core entries, but this time I also followed any index entry for "location", which is Britannica's way of taking you to a latitude and longitude table.

I found entries for all my candidate mountains in both sources, although Chances Peak appeared under the (not quite identical) guise of "Soufriere Hills Volcano" in Wikipedia, and I had to recognise that "Yding Forest Hill" was Britannica's rendering of Yding Skovhøj.

Heights were provided for 19 mountains in Wikipedia, and all 20 in Britannica. Both sources were out of date for three mountains (Kilimanjaro, Vinson and Gunnbjørn), and made clear mistakes on one each - Olympus for Wikipedia and (sigh) Ben Nevis for Britannica (1343m / 4406ft; 1344m/4409ft is correct). Wikipedia also contained one error in converting from metres to feet. Britannica gave latitudes and longitudes for all summits but Chances Peak; Wikipedia managed only 14 latitudes and longitudes, and made small errors in two of those it did give. One of the lat/long pairs missing from Wikipedia was for Ben Nevis - instead, the entry provided a quirkily UK-centric Ordnance Survey grid reference.

Britannica was hugely better on providing geographical detail, and noticeably so with regard to geology and history. Wikipedia had the better discussion of names (it recorded the 1997 spelling change of Kosciuszko, while Britannica still used "Kosciusko"), and it was much better on ascent routes and the details of first climbs. It also provided images for more than three-quarters of my mountains (albeit with two mislabelled), whereas Britannica managed to illustrate only half of them. Neither source was much interested in biology, with Britannica's enthusiasm for the vegetation zones of Kilimanjaro representing only a brief efflorescence in an otherwise pretty dry season.

image from The Angry Corrie

Britannica's standard of writing was much the higher, as you might expect for paid communicators. Wikipedia was patchy in style and structure, and in my 20 entries I found two sentences I actually couldn't understand.

It was disappointing, in my small test, to find that Wikipedia's height data were no more up-to-date than a three-year-old paper encyclopaedia - online information providers have been announcing the death of paper reference for years, simply because the publishing delay of the latter is so long, but the opportunity to kick laggardly Britannica ass was almost entirely missed.

Yeah but, the Wikipedians cry from the sidelines: Wikipedia is FREE. That's true. It's a free source of knowledge written by enthusiasts, with some very good bits and some not very good bits, and it can be both an education and a frustration. But that's just the internet, that is, gently strained of its lunacy and pornography. If you're knowledgeable about a topic, chances are you've already got a tried and trusted reference source you'll use in preference to Wikipedia; if you're not knowledgeable, chances are you'll do an internet search for something specialised, or at least to cross-check what Wikipedia offers.

Which does leave me wondering a bit about the point of Wikipedia.

TAC 65 Index