A lot has been written about the future of Wikipedia. Given some recent involvement I have had with the Wikipedia deletion process, it set me to thinking about who determines what is important and what is not on the Intenet. Read on for my conclusions...
Our Modern Salon Des Refuses
In 1863, Napolean III ordered the first Salon Des Refuses (Salon of the Rejected) be held in Paris. The goal of the Refuses was to showcase the increasing number of submissions refused from the main Parisian Salon. The traditional Salon had grown greatly in its notoriety, and along with it the number of submissions had swelled. Quickly the jury-administered submission process found itself turning out more rejections than acceptances. Napoleon, a bit anxious over the increasing public perception that there was a cultural “tyranny of the few” built into the jury selection process, headed the issue off at the pass by introducing the Refuses. In retrospect this was a brilliant maneuver. Some exhibitors in the 1863 Salon Des Refuses included such complete artistic hacks as Paul Cézanne, Camille Pissarro, Armand Guillaumin, Johan Jongkind, Henri Fantin-Latour, James Whistler, and Édouard Manet. Yeah, yeah, I know – I have never heard of them either?
Scanning a more contemporary horizon, we can find a similar situation emerging in our own technical back yard. Due to its immense popularity and singularity of distinction, the number of submissions to Wikipedia is skyrocketing on a daily basis. As with all collections that depend (and are valued) to some extent on the perception of quality, with the influx of so much new content, Wikipedia has ratcheted up the number of administrators (e.g. the jury) it uses to select which submissions are worthy and which are not. In fact, there was a much publicized debate over Nature’s special report that Wikipedia’s quality was approaching that of the Encyclopedia Britannica. The net/net of it all being a sort of “horse shoes, hand grenades, and Wikipedia – its close enough” kind of thing.
To look at some numbers, let’s consider (as of this writing) that 995 administrators exist to monitor roughly 1,336,000 valid articles (English article count only) with a mean of 35 edits per article. Doing some rough math this means each of the 995 administrators handle the monitoring of 1342 articles, delete an article every 4.7 days (although it usually happens much faster), and assuming all edits to an article occur in a 12 month period (which is probably a bad assumption) – review 129 edits per day. By no means is this a light load and a moment of thanks must be taken to acknowledge this voluminous volunteer effort. But in the end, the only approach that can be taken is Wikipedian tough love. Compare for example the daily average of 2106 new articles per day in July to the 209 article deletions that happened just yesterday. At a 10:1 ratio this starts to feel like a pretty quick and dirty approach to picking the right apples from the cart. And to some extent if you survey the Netgeist, one gets the feeling that we may be leading up to our own replay of the Salon Des Refuses pageant.
Back to Reality
With these stats, one could argue that it might lead to some quick and not always correct decisions on the part of jurors. If Whistler can get glossed over, I’m guessing a retired university professor in Sheboygan might miss the relevancy of something related to ERP systems. To this end, the statistic that strikes me as most important to consider is the 1:10 submission to deletion rate. Anecdotally and personally this is not far from my own experience. What’s interesting is that a lot of what I submit is deleted for one of the following reasons:
1) It’s a neologism (I am trying to go down in history here people!!)
2) Commercial (heaven forbid)
3) Of dubious utility (which can mean a few things)
It’s actually not the deletion that frustrates me (I understand that a lot of crud gets tossed into the soup), its that fact that every article that is rejected currently is not returned to the original author for follow up. As the author, you have to hunt down the deletion and kow-tow a bit to the administrator to argue your case. What’s annoying about this is that most administrators actually do put a bit of effort into validating the “utility” of the entry by Googling around, searching Amazon, etc.. Basic respectable web research. So time is spent and it’s not totally haphazard. This is a good thing! However, I am reminded of a quote I hold near and dear from my time at MIT, “How do you get the right answer to a question? Ask someone who knows it.” The point being that one of the flaws in Wikipedia’s submission process is that the original submitter is not directly included in the jury review process for deletion of an article.
There are actually some fantastic and well documented examples of this. A recent one occurred with my fellow Irregulars when one of us submitted an entry for “Enterprise 2.0” to Wikipedia. It was quickly deleted by an administrator for all of the above reasons. After a long and vocal few days, some blog flogging, and finally a gentle nudge from some Wikipedia insiders (“honey catches more flies than vinegar”, me thinks) we were finally able to resuscitate it. The saga showed just how flawed the jury process can be.
Another humorous but stark example involved an administrator insisting on the removal of edits to an article on Knowledge Management because he felt a notable academic and de facto co-founder of the KM movement would completely disagree. The person adding the originally deleted comments was Dave Snowden, that very academic himself.
These pithy examples set aside, there is clearly an emerging dialog around what sometimes is perceived as the tyranny of the few (Wikipedia administrators) selecting what is relevant to the world from an information perspective.
It’s rare that I actually get to use the actual word parallax in my blarticles. There are in fact quite a few derivative meanings of this word. My blog was actually named after one specific definition: the geometric calculation of distance to an object using only a couple known points of reference. There is however another definition of parallax which is more frequently referenced in video game design and astronomy and can be used to describe the Wikipedia problem relatively well.
In old video games, rather than have completely fluid three dimensional models, a series of layers would be created in the background. These layers would then be scrolled at different horizontal speeds as the character moved. This parallax (or motion parallax) would give the appearance of three-dimensionality at a very cheap computing cost. Another way to think of this at a more metaphorical level is that the view of an object changes depending on the background or context given to it (something very detailed against something not very detailed would appear to be in the foreground).
Here is a visual example of classic parallax (from Wikipedia of course):
Getting back to Wikipedia and the nature of information, one realizes that we’re dealing with a similar type of parallax situation. Take for example myself, Niel Robertson, as a subject. As far as a Wikipedia administrator would care, an article on me would be deleted because I have “dubious utility” (just some random some tech entrepreneur with little distinction). If all of a sudden we shift our context to Newmerix, my utility becomes arguably less dubious (yes – some, maybe many, would disagree with that statement). The point being that given different contexts the perspective on information’s value changes. I call this phenomenon “Information Parallax” (a quick Google search and I think I might actually have coined this phrase – wahoo – let the legacy begin).
Information Parallax: the apparent change in utility of a piece of information based on a change in the context in which it is considered.
And herein lays the problem for Wikipedia. And in exactly the same way that faced the original Salon’s jury. Some notes on early entrants into the Salon Des Refuses (for an in depth discussion check out the Judgment of Paris) have claimed that they were rejected from the main Salon as they were perceived as offensive or inconsistent with classic styles. It would be fair to assume then that the context of the main Salon was neither pornography nor experimental works. While it’s hard to imagine Cezanne or Whistler being accused of being either now-a-days, lets not forget that people rioted when they first heard Stravinsky’s Rite of Spring). Change the context to contemporary views on sexuality and what might be defined as art and we clearly see Whistler et al drawing well within the lines.
Oddly enough, what is happening in our modern times is not so different than what happened in Paris. Clearly our Wikipedian Napoleon, Jimmy Wales, hearing the public murmurs, saw the need, and headed this issue off at the pass. To this end, Jimmy et al at Wikipedia.org have set up the Wikimedia Foundation to capture certain non-Wikipedia types of content such as books, species information, and dictionary entries. Our very first official Salon Des Refuses!
Other non-Wikimedia related “Refupedias” are starting to pop up as well. I recently discovered Yellowiki which provides wiki-based open descriptions of local businesses. In the context of Wikipedia classic, any entry in Yellowiki would be deleted immediately by a Wikipedia administrator, but in the context (information parallax) of local businesses, the information is highly useful. For the very reasons it might be deleted from Wikipedia (“too commercial”), it provides utility to a consumer in a competitive market. Another similar example would be ShopWiki which contains store and product listings. And of course you can now set up your very own SocialText wiki or jump in and contribute to a more specifically focused one that someone else has created.
Another fantastic example which demonstrates the immense power of taking a different view on content that Wikipedia is WikiTravel. WikiTravel is the leading wiki designed for collaborative management of “Lonely Planet” style travel entries. While doing some research for a trip to Iceland, I actually thought of this idea myself. Using the Lazy Web, I was not surprised to find WikiTravel already pretty far down the path. Again, a standard WikiTravel page would surely be rejected by the Wikipedia submission jury because it contained the names, phone numbers, and, say, day rates for bike tours around Reykjavik.
However, there is another reason why I am coming to love WikiTravel. One of my frustrations with any guide book that you buy is that 1/3 of the information by definition is outdated by the time you buy the book. Have you ever had success looking in a Rough Guide or, heaven forbid, a Fodors trying to find the hottest night club to go to on a Thursday night in the Zona Sul in Rio? Well even if your version of adventure travel isn’t quite that exotic - the general half life of the hottest nightclub, lounge, Broadway show, or restaurant in New York is still probably 6 months. The overall point being that the accurate answer to the question actually changes relatively frequently. And WikiTravel does a fantastic job of demonstrating a seldom heralded value of wikis – timeliness and frequency of change. End the end, the nature of a wiki is perfectly related to solving this problem, and now with WikiTravel, the content itself can have a friendly home.
To some extent I think we’re going to see a sort of Peak Oil for Wikipedia submissions. As Wikipedia moves to a maturity phase, Jimmy’s new mantra has already turned into their “quality initiative”. Once this kicks into full gear, there will be a steady decline in the number of new article submissions and a steady increase in the number of edits per article (which Jimmy correlates roughly to quality). I am not sure where the peak comes but I am guessing its around June 17th 2008 (completely unsubstantiated predictions are nothing without our extremely specific dates!).
However, much like what’s occurring with alternative energy sources now, if you believe in the Information Parallax concept, you’ll probably agree that we’ll see a number of large wikis emerge a step of two further down the semantic tree to provide a home for all Wikipedia rejects. Giving a Yellowiki entry the context of “Local Business” and the WikiTravel entry the context of “Travel” massively shifts value of its content.
Don’t get me wrong, there is immense value in Wikipedia and the path that it is heading down. A jury based process should theoretically enforce a higher level of quality. However, history has shown us that subjectivity is always, well, subjective. Apparently we need our Salon Des Refuses just as much as the Parisians did.