My wife and daughter are going on vacation soon, and so for one week, I'll be able to fully indulge my natural inclination to stay on the net 18 hours a day. :-) So I plan to do some things that I haven't had time for.
Once upon a time, I did a study of our site and Britannica's, comparing our articles to theirs. As I recall, I used our 'random article' feature to select ten topics, and then I compared theirs and ours. (But I seem to remember finding some subjects on which we didn't have an article and they did, so I don't know how I could have done that.)
What I'd like to do is sit and explore Britannica and Wikipedia, comparing and contrasting our strengths and weaknesses, with an eye towards my REAL project for the week, which is to put together a "Roadmap to 1.0" for Wikipedia.
So far, we have put aside most talk of 'finalizing' articles for a 1.0 release because there was so much left to do. But now I think that we have a huge number of articles, nearly complete coverage in many common areas, and that it would be wise to start thinking about how and when we might produce a '1.0' version to make it easy for content re-users to produce CD-ROMs, books, whatever.
One of the crucial components in deciding when to make a push for something like that is our overall level of completion. And measuring against Britannica is a good way to assess that.
What methods should I use? I obviously can't read both works cover to cover in a week. :-) So I have to randomly sample somehow, I guess.
And I might supplement my random sampling with some 'linear' checks like: do we have an article on every leader of every nation? Do we have an article on every city in a list of the 100 most populus in the US? Things like that.
--Jimbo
Jimmy-
What I'd like to do is sit and explore Britannica and Wikipedia, comparing and contrasting our strengths and weaknesses, with an eye towards my REAL project for the week, which is to put together a "Roadmap to 1.0" for Wikipedia.
In my opinion, Wikipedia should be the unstable version of Nupedia, a bit like Debian has a permanent unstable section. Things in Wikipedia can be complete nonsense, but they are always up to date. We may add a team certification model to Wikipedia eventually, but I'd be happy to see a simple Sifter solution as envisioned by Magnus (without the limit to certified experts as reviewers which I think is what Larry wanted). IMHO it would make perfect sense to use the Nupedia name and domain for that project.
Nevertheless, a "State of Wikipedia" document would be useful. We should create this for the English Wikipedia specifically, and describe reviewing methodologies. I've done such reviews myself, first with Microsoft Encarta and recently with Britannica. In both cases, Wikipedia often had much more and more accurate material. It suffers especially in the foreign politics department for countries that few people know about. Much of our information there is still the CIA text. Our history section has substantial gaps or contains old Britannica content, but we have many, many history buffs who are very active correcting this.
A good approach to see where we have weaknesses is to take a look at the "Oldest articles" list (from the Special pages menu). These are the ones that people apparently don't look at, or cannot contribute anything to. It would be useful to compare the quality of those to the quality of articles in another encyclopedia.
Eventually we may want to create a [[Wikipedia:WikiProject Patchwork]] specifically to address our weaknesses. This would not be a "let's get ready for 1.0" thing, but a permanent improval process. The 1.0 release for CD-ROMs etc. would be prepared by the Nupedia editors.
Regards,
Erik