Hacker News Reader

Comment by nkozyra | Hacker News Reader

Attacking Wikipedia when the citations for the claims are very clearly stated within is tiresome. If you'd clicked on the citations within you'd find:

American Journal of Physiology, The American Journal of Clinical Nutrition, Annals of nutrition & metabolism, etc.

If you have issues with those sources, fine, but saying "don't cite Wikipedia" when the article is well-cited is not useful in the least.

View on HN

Thank you for that clarification. My point was that basing your argument on a paragraph from the Wiki doesn't really hold water. Have you gone through those cited articles? Have you checked the data? You trust that those statements next to the citation are representative of what is actually been said in the article itself. Wikipedia is great, I am all for it. I just personally wouldn't building my argument solely based on what is said there.

The scope and depth of human knowledge is so large that validating accuracy of information is humanly impossible for most interesting subjects beyond experts and the very interested.

You're absolutely correct that information needs to be validated, but wrong in singling out Wikipedia. This rabbit hole goes pretty far. I've found false citations from published articles. Conservatively assuming 25 citations per article, the amount you have to validate is 25^(citation depth).

There are 11 citations in the above quote, so here's the homework you're requesting: Level 0 validation: 11 Articles Level 1 validation: 275 Articles Level 2 validation: 6875 Articles

The knee-jerk distrust of Wikipedia is so old that a paper was published in Nature in 2005 showing accuracy close to that of the Encyclopedia Britannica.

https://en.wikipedia.org/wiki/Reliability_of_Wikipedia

Just like any other source, you have to assess it. How many editors, to which degree is it a target for vandalism or astroturfing, how well is the article written, the type of the subject material, etc.

Therefore I would expect britannica-level accuracy on the caffeine page, and that's pretty good for most discussions.

Fair enough. Valid points. Still I believe we conflating things here. I believe encyclopedias are good source for well established facts, don't think they are the main source for citation when discussing hypotheses that still undergoing testing. When it comes to Wikipedia, you basically trust it to be up to date with state of the art. Yes, research seems to show that it converges towards some notion of the truth, but it takes time and ultimately you don't know if what you are reading at the moment is a true representation of what was actually written in the original paper. Again, not saying this is bad, I would just be cautious about accepting an argument which is solely a copy&paste from a Wikipedia page that makes claims based on handful publications that I presume the person posting haven't read. To sum up, I am not against Wikipedia, I am somewhat cautions accepting wild statements like "Drinking coffee can Kill you" just because someone took the time to put it into a wikipage (with a bunch of citations).

http://www.9news.com.au/national/2015/09/15/12/13/australias...

Even wikipedia tells people not to use wikipedia as a source. It's a reasonable point.

Reply on news.ycombinator.com