5 ms·

One of the major argument I have seen why the retention graph has gone down is that English Wikipedia is much more complete in 2021 than 2007. Writing about eso

by belorn 5y ago

One of the major argument I have seen why the retention graph has gone down is that English Wikipedia is much more complete in 2021 than 2007. Writing about esoteric topics is by itself limiting to fewer editors, and the bar of notability is more restrictive. Non-English articles is easier in that sense, but much of that work seems to become a translation of the English version. Writing a non-English article from scratch is an alternative but can feel a bit less interesting when the English version is already of high quality. I could be wrong but today it seems that most editor efforts in the English Wikipedia goes towards news, politics and media. It would not surprise me if those have a natural lower retention rate than other form of articles.

strogonoff 5y ago
On one hand, Wikipedia has become more “complete” in the last 10 years (though information never is, really). On the other hand, the grown amount of content means there is proportionately more content to maintain—which might take even more effort and attention to detail than writing new content. That, as well as the reputation and pervasiveness of links to Wikipedia, together make it more vulnerable to and a more compelling target for instances of blatant or stealthy misinformation, censorship and vandalism. On that topic, I suspect a resource that tracks change histories for sensitive Wikipedia articles[0], aggregating and correlating changes by IPs and usernames along with some smart NLP analysis on text contents of the diffs, could end up being insightful. [0] Many of us know a certain date coming up, publicly remembering which starting this year is already punishable by a prison sentence in Hong Kong. Maintaining the article covering this one is hopefully taken care of for now as it is pretty conspicuous—but for how long, and how many lesser known articles are there…

by belorn 5y ago

strogonoff 5y ago

On one hand, Wikipedia has become more “complete” in the last 10 years (though information never is, really). On the other hand, the grown amount of content means there is proportionately more content to maintain—which might take even more effort and attention to detail than writing new content. That, as well as the reputation and pervasiveness of links to Wikipedia, together make it more vulnerable to and a more compelling target for instances of blatant or stealthy misinformation, censorship and vandalism. On that topic, I suspect a resource that tracks change histories for sensitive Wikipedia articles[0], aggregating and correlating changes by IPs and usernames along with some smart NLP analysis on text contents of the diffs, could end up being insightful. [0] Many of us know a certain date coming up, publicly remembering which starting this year is already punishable by a prison sentence in Hong Kong. Maintaining the article covering this one is hopefully taken care of for now as it is pretty conspicuous—but for how long, and how many lesser known articles are there…