this blog is

17 November 2004





Wikipedia logoThanks to Chris (no, not that one, or that one either!), who pointed me to Robert McHenry's criticism of Wikipedia, entitled the Faith-based Encyclopedia.

Firstly a general thought, not related to McHenry's article specifically: The word "encyclopedia" carries with it associations of authority, fairness, accuracy, comprehensivity and so on that the wikipedia can't always live up to for all topics at any given time. So whilst the term does encourage a high standard for contributions, it also sets a high standard for comparison. However, ignore the connotations with what we usually regard as an encyclopedia and you still end up with a very useful resource. One that is, in fairness, generally far more reliable, comprehensive, accurate, etc than the surrounding internet. So, I use wikipedia, but I just pretend the "-pedia" suffix isn't there.

Getting back to the McHenry article, the main point of which is to criticise the collaborative model for wikipedia:

Some unspecified quasi-Darwinian process will assure that those writings and editings by contributors of greatest expertise will survive; articles will eventually reach a steady state that corresponds to the highest degree of accuracy.

He claims that there is no basis for believing this statement, hence faith is required.

It would not be hard to disprove this, simply by finding an article that had achieved the required standard through repeated editing and refinement. Maybe the cited Alexander Hamilton article does not meet this criteria (though see below), but others almost certainly do. Take a less obscure, non-controversial topic like, say, the Dalai Lama and you find a good basis for believing in the collaborative editing model used by wikipedia.

So the characterisation of (part of) the wikipedia editing model as "entirely faith-based" is not fair in my opinion. There is supporting evidence in favour of the wikipedia model, hence faith is not required in all cases.

Given that it is easy to see how the wikipedia model would work for certain topics, it is also easy to see how it would not work for other, more obscure or more controversial, topics. However, contrary to intuition, it often does work for such topics. Tim Lambert provides some evidence for this, based on his experience with the entry for pro-gun advocate John Lott. Although it seems that Lott himself has taken to removing parts of the wikipedia article that are criticial of him, "the loudest voice has not won".

Returning the example of Alexander Hamilton cited in McHenry's article. Look at the entry now: it does list the two possible birth years, and the fact that there is some controversy. Meaning that when the article was written the wikipedia process had not yet achieved the "steady state that corresponds to the highest degree of accuracy". In other words it is hardly fair to pick a single entry and say that it has not achieved the desired standard, when the process clearly allows for further refinement. Admittedly these changes may have been triggered by McHenry's article - but the point is that they were made.

Likewise it is hardly fair to use this example as the basis for the claim that that the article has been "edited into mediocrity", as there are no guarantees that each successive revision will be an improvement on the previous one. If you only pick one snapshot of one article you are unlikely to observe useful trends, and any conclusions you draw about either the entirety of the wikipedia, or the process used to generate it, are immediately suspect.

I am not at all convinced about the authors guess at the inspriation for the model for the as based on "journaling" (which I must admit I have never even heard of - maybe a US educational policy?). It doesn't even sound like a likely explanation to me. I suspect wikipedia is more inspired by collaborative development models for open source software than anything else.

I take McHenry's points about not being friendly to encyclopedia users: those who just want to find the information they are looking for and go away. But he's missing the point: wikipedia really isn't that kind of resource. TANSTAAFL, after all. Part of the cost to the user for accessing this excellent resource is the responsibility to a) assess the information presented in a critical manner, given the vagueries of the editing process, and b) contribute something back when you can.

On a slightly snarky note, Lambert has in the past made some observations about the quality of posts on the Tech Central Station: "Sometimes I think that there must be a qualifying exam in order to write for Tech Central Station. Fail the exam and you’re in." Well, it's certainly more deterministic than a faith-based model...


Posted by
2004-11-17 11:23:39 -0600

I read this article too; you've saved me (more of) a rant. I reckon Wikipedia is great for totally uncontroversial entries and I actually think it's pretty good for very controversial entries, because of the way it alerts you to claims of bias and sends you off to the discussion page (which admittedly does often break down into unhelpful sniping -- look up Palestine sometime). I think there's a kind of hazy mid-range though of entries that aren't exactly clear-cut but which not many people know or care much about, so they don't register as controversies and dodgy information goes unchallenged (which category thingo's entry may fall into). Overall though I think it's a great resource and use it a lot; it'll only get better as more information is digitised and made available and more wonks become aware of Wikipedia; and I reckon it will keep evolving meta-processes to ensure that the editorial quality ratchets up and not down. Finally, thingo writes for a traditional encyclopaedia; and every encyclopaedia has a certain amount of rubbish in it, which tends not to get corrected for years if ever.

Posted by
Peter Marks
2004-11-17 11:23:39 -0600

Wikipedia may not be perfect but it is amazing how much good content is there. It's a better place to get an overview of a subject than just the "I feel lucky" link in google.

I put phpwiki in at the office, updates are recorded by IP address so if required we can see who did what. It's a huge hit. All our procedures, notes on the network, project notes, and all sorts of random stuff that would normally be locked away in documents on an individual's computers are shared there.

I bought the encyclopaedia britannica in book form (!) from a local library for just $120. It's a lovely thing to flip through and own, but day to day we all use the net to find out stuff.