On Fri, Mar 7, 2008 at 6:35 PM, Ian Woollard ian.woollard@gmail.com wrote:
Probably not. The thing is the wikipedia gets to be the top of google searches because it's generally fairly reliable. Likewise high up in the web rankings. If we start allowing less obviously notable things in, then the average quality can only go down, and eventually that will get reflected in how people treat us.
I'm afraid this is an example of rather muddy-headed thinking on your part.
Firstly, we honestly should not give a damn whether our Google rankings are high or low. We're not in the business of producing Google rankings, and we don't get ad revenues from the hits.
Secondly, our Google rankings and our rankings on peoples' personal indicators of reliability are to do with the fact that we are likely to have pretty good content on a topic searched for, compared to what else is available online in one place.
If a person searches for something, the existence or non-existence of articles unrelated to their interests doesn't alter the quality of the information they find on what they ARE searching for. It's simply invisible.
However, if someone searches for something and we DON'T cover it - because some ad hoc 'consensus' of half-a-dozen AFD regulars decided it was 'non-notable' - then Wikipedia has failed to provide information. That tends to leave people with a bad taste in their mouths; double that if when they searched last week, Wikipedia DID have the information they wanted and now doesn't.
'Non-notable' content only 'hurts' our quality if you are measuring quality by some pretty odd measures.
Yes, it's embarassing when some blogger points out that we cover some random Pokemon better than we cover a head of state. The solution to that isn't to delete the Pokemon.
I think this is one way in which we are - as we often are - WAY too over-sensitive to outside criticism. We handle it poorly and panicked, and often make matters worse.
Yes, there's always going to be losers with any thing that helps maintain quality. Most of the measures we have are only correlated with quality, not direct measures. That means that some things get excluded when they shouldn't. But my point is that this is probably unavoidable if we want a high quality encyclopedia, which to be brutally honest, it doesn't sound like you do, particularly- you value covering 'everything' higher.
It might be better if you didn't assume that because people disagree with you they don't value the encyclopedia.
I also see no reason why we have to use any automated measure of quality that creates 'losers'. We don't have to do everything by algorithm.
-Matt