I thought Google will result in 90% of Wikipedia page views ;)
...or maybe that's correct but the stats are for the whole of Wikimedia grouped together.
I wonder if the other services/sites of wikimedia as much "popular" as wikipedia though.
depends on how you count. If a person googles for "SEO", klicks on en.wikipedia.org/wiki/SEO and from there to en.wikipedia.org/wiki/Search_e ... the statistic would indicate that 50 per cent of all page requests came from google, not 100%. There is no tracking of sessions that might have started at google but ended up in two hours of wild clicking arond wikipedia.
And on the other hand, many people I guess are like me and use Google as "Wikipedia search engine". So I *want* to go to Wikipedia because I suspect it has the best article on something, but I first go to Google to then search for [keyword wikipedia]. So in that case, Google is logged as the referrer but it's not like it brings people who'd otherwise not have ended up on Wikipedia. I guess if the Wikipedia search engine would be better, people would also go less to Google as a by-stop towards the Wikipedia article...
Philipp: Over the past weeks, I think I saw some great improvements at ls2.wikimedia.org. What do you think about the search results there?
They're probably better, but I think it's hard to beat Google. For instance, when I want to look up the article on Wikipedia describing the history of the Israel/ Palestine war I can type something like [middle east conflict wikipedia] into Google... and the top results will be
Arab-Israeli conflict – Wikipedia, the free encyclopedia
Some uses of the term Middle East conflict refer to this matter; .... Find more information on Arab-Israeli conflict by searching Wikipedia's sister ...
Israeli-Palestinian conflict – Wikipedia, the free encyclopedia
From Wikipedia, the free encyclopedia. Jump to: navigation, search .....
The Complete Idiot's Guide to Middle East Conflict. 2nd ed ...
It's "unfair" when I say, yes, the second one was exactly what I meant/ wanted to pull out – because I "should" have entered "israeli-palestinian conflict" into Google, or something along the lines – but somehow what's important is that Google understands how many backlinks point to a respective article, thus giving me the most important first. ls2.wikimedia.org on the other hand for the same search, [middle east conflict], returns these (possibly more literal) results:
1.00 – 2006 Lebanon War
0.98 – Conflict: Middle East Political Simulator
0.85 – List of conflicts in the Middle East
0.72 – Middle East
0.67 – Middle East Forum
0.65 – Arab-Israeli conflict
0.65 – American intervention in the Middle East
0.62 – Democracy in the Middle East
0.62 – The Complete Idiot's Guide to...
0.61 – History of the Middle East
0.61 – Anthony Cordesman
0.61 – Barry Rubin
0.61 – Bernard Lewis
0.61 – Rashid Khalidi
0.60 – Conflict
0.60 – International reactions to the 2006 Lebanon War
0.60 – Etiquette in the Middle East
0.60 – Russia and the Arab-Israeli conflict
0.60 – Conflict (disambiguation)
0.59 – Media coverage of the Israeli-Palestinian conflict
Here's another example; I enter [google hoax wikipedia] into Google, and get:
1. Google's hoaxes
2. Google juice
The first one seems like the perfect result for my query – a page covering all kinds of hoaxes. ls2.wikimedia.org on the other hand returns:
1. Google bomb
2. Jew Watch
2. April Fools' Day
3. Bogdanov Affair
4. April 1, 2005
5. April 1, 2007
So these range from things where I can't see a direct connection (Bogdanov Affair) to things which are related, but not strongly enough (April 1st is April Fool's day, but it's not a page all about Google) to individual Google bombs.... but the best page, "Google's hoaxes", is missing. Now of course it's a matter of taste what "best page" is but I'd make a bet you could do some user blind test and Google would score higher satisfaction results...
"Wikipedia has the best article" ... sorry, you lost me there. Without any professional credibility or quality control, Wikipedia hardly has the "best article" for anything. Each article needs to be vetted against more credible resources.
Thanks for reviewing ls2 :)
About the first search, you are right, google does a much better job of getting the most relevant article. This is possibly because they have a more elaborate scoring mechanism and page links analysis.. LS2 page ranking is somewhat simplistic, and is only there to filter out most of the irrelevant pages (some still remain :( ). Hopefully, at some point it will get better :)
About the second query, I think you misinterpretted the results. You entered "google hoax wikipedia" into LS2 search, and there is no reason to give you google's hoaxes as first result, as there is no mention of wikipedia in the article – i.e. all the words you enter in the query are required to be in the article or in the title ... and google hoaxes doesn't have anything to do with wikipedia... if you enter "google hoax" you will get the google's hoaxes as the first result... As far as I know, Google found those pages with "google hoax wikipedia" only because it indexes the whole page, not only the contents..
> "Wikipedia has the best article" ... sorry, you lost me
> there. Without any professional credibility or quality
> control, Wikipedia hardly has the "best article" for anything.
Who are you quoting Michael?
In regards to what my comment, I merely commented on the *best page on Wikipedia* on the subject (because we are discussing the relevancy of results only in the context of Wikipedia site search).
> About the second query, I think you misinterpretted
> the results.
Sorry Robert, I entered a wrong search – I take back my second example, I obviously wanted to enter [google hoax] into the Wikipedia search (not [google hoax wikipedia]). And doing so, it actually does return the page I was going for...