Sunday, January 07, 2007
Writing articles, part III
This is not actually about writing new articles, but helping to improve existing ones, like Jimbo stood up and asked all of us to.
One of the things that impressed me about Jared Diamond's Guns, Germs, and Steel was not his thesis, but the bibliography in the back of his book: it was comprehensive, just like every article on Wikipedia ideally should be. And after reading Diamond's book, I was tempted to create a little disruption by comparing that bibliography with the references section at the end of several Wikipedia articles. My point was simple: when we havea source like this to mine an example like this to use, why shouldn't every related article be that much better?
However, after having a look at several articles on topics related to the ones in diamond's book (specificly, domesticated plants and animals), I felt that they weren't that bad. Many, I felt, were perfectly acceptible for inclusion in a proprietary print encyclopedia. I guess this shows just how fast one's impressions of Wikipedia can become outdated.
Diamond's bibliography alerted me to a very useful book: Daniel Zohary and Maria Hopf's Domestication of Plants in the Old World. At the county library I found a copy of the third edition, and after a quick review of it I decided to start adding information from their book to relevant Wikipedia articles. After all, one weakness in many articles is not the presence of incorrect information (which can always be removed), but the presence of correct (or mostly correct) information for which there is no source. Adding information from Zohary and Hopf's book would would mean that I only needed to insert a sentence or two with a footnote to about 50-60 articles. I estimated that this would take about 10 uninterrupted hours of work -- at most a week of start and stop efforts. Each of these articles were likely well written, and either had a consistent system of citations -- or none at all; the barrier to entry for this mini-project should be quite low, I reasoned.
What I found was an undetected, but very high barrier, to this plan: for some articles, I had to figure out just how to insert a comment about the domestication of a given species or genus of plant in the article so that it would not break the flow of the narrative. In others, the citatino system was so broken that I could not figure out how to use it. After about a dozen articles, I called it a day, and decided to help Yvette take down the Christmas tree instead.
I would be easy to blame this problem on other Wikipedians, because every article should be written to be as edit-friendly as possible. But there are only a few thousand active Wikipedians to shepherd over 1.5 million articles, and often it is all an editor can do to put some material into an article and hope that the next person along will fix it -- or to simply flag an article that feels wrong, but one lacks the knowledge to actually diagnose and fix it. The system of WikiProjects should prevent this, but the active ones are still too few, and the majority of them barely function.
I honestly wish I had a solution to fix this, to make Wikipedia better in this regard -- besides the obvious solution of lots of hard work. My current tactic -- write the article in a structured, edit-friendly form from the beginning -- in a way defeats itself, because every article in Wikipedia arguably should be flexible to change. (And some might not like the structure I pick for my articles.) The solution that one should simply accept the fact that Wikipedia will always be uneven in quality might be more realistic -- but again fails to solve the problem: we accept second-best when Wikipedians are being asked to strive for the first place.
At least by acknowledging that there is a problem, I hope that someone smarter than me knows the problem exists, examines it, and will offer a solution.
Geoff
One of the things that impressed me about Jared Diamond's Guns, Germs, and Steel was not his thesis, but the bibliography in the back of his book: it was comprehensive, just like every article on Wikipedia ideally should be. And after reading Diamond's book, I was tempted to create a little disruption by comparing that bibliography with the references section at the end of several Wikipedia articles. My point was simple: when we have
However, after having a look at several articles on topics related to the ones in diamond's book (specificly, domesticated plants and animals), I felt that they weren't that bad. Many, I felt, were perfectly acceptible for inclusion in a proprietary print encyclopedia. I guess this shows just how fast one's impressions of Wikipedia can become outdated.
Diamond's bibliography alerted me to a very useful book: Daniel Zohary and Maria Hopf's Domestication of Plants in the Old World. At the county library I found a copy of the third edition, and after a quick review of it I decided to start adding information from their book to relevant Wikipedia articles. After all, one weakness in many articles is not the presence of incorrect information (which can always be removed), but the presence of correct (or mostly correct) information for which there is no source. Adding information from Zohary and Hopf's book would would mean that I only needed to insert a sentence or two with a footnote to about 50-60 articles. I estimated that this would take about 10 uninterrupted hours of work -- at most a week of start and stop efforts. Each of these articles were likely well written, and either had a consistent system of citations -- or none at all; the barrier to entry for this mini-project should be quite low, I reasoned.
What I found was an undetected, but very high barrier, to this plan: for some articles, I had to figure out just how to insert a comment about the domestication of a given species or genus of plant in the article so that it would not break the flow of the narrative. In others, the citatino system was so broken that I could not figure out how to use it. After about a dozen articles, I called it a day, and decided to help Yvette take down the Christmas tree instead.
I would be easy to blame this problem on other Wikipedians, because every article should be written to be as edit-friendly as possible. But there are only a few thousand active Wikipedians to shepherd over 1.5 million articles, and often it is all an editor can do to put some material into an article and hope that the next person along will fix it -- or to simply flag an article that feels wrong, but one lacks the knowledge to actually diagnose and fix it. The system of WikiProjects should prevent this, but the active ones are still too few, and the majority of them barely function.
I honestly wish I had a solution to fix this, to make Wikipedia better in this regard -- besides the obvious solution of lots of hard work. My current tactic -- write the article in a structured, edit-friendly form from the beginning -- in a way defeats itself, because every article in Wikipedia arguably should be flexible to change. (And some might not like the structure I pick for my articles.) The solution that one should simply accept the fact that Wikipedia will always be uneven in quality might be more realistic -- but again fails to solve the problem: we accept second-best when Wikipedians are being asked to strive for the first place.
At least by acknowledging that there is a problem, I hope that someone smarter than me knows the problem exists, examines it, and will offer a solution.
Geoff
