on 8/8/08 10:03 AM, David Gerard at dgerard@gmail.com wrote:
Here's some numbers: http://oc-co.org/?p=124
It was I who brought up the subject most recently, David. I was presenting to those who say that Wikipedia is great, the best, because of its size. And the emphasis on edit count as a criterion for adminship (among other things) - again numbers, size. My suggestion was to spend more time on improving the quality of the Articles that exist now.
Marc Riddell
On Fri, Aug 8, 2008 at 10:30 AM, Marc Riddell michaeldavid86@comcast.netwrote:
My suggestion was to spend more time on improving the quality of the Articles that exist now.
So your suggestion to improve quality is quantity (of time spent).
While we're making useless suggestions, my suggestion is to spend the same amount of time on improving the quality of the Articles that exist now, but to do so more efficiently. Or even better, my suggestion is to invent an article-writing robot to automatically write a perfect article if we only feed it an article title.
Here's some numbers: http://oc-co.org/?p=124
- d.
Yes, a good thing really to have plateaued in this way. Much remains to be done however. I recently imported the article "Tasiilaq" to Wikinfo and was quite disappointed with the lack of detail about the city. It is not at all clear how anyone manages to make a living there.
Fred
2008/8/8 Fred Bauder fredbaud@fairpoint.net:
Here's some numbers: http://oc-co.org/?p=124
- d.
Yes, a good thing really to have plateaued in this way.
Not really. While a case could be made for an article plateau (at around 100 million articles estimates vary) not for an edit rate plateau and even there we would expect growth at a level roughly equal to web growth rates.
2008/8/8 geni geniice@gmail.com:
Not really. While a case could be made for an article plateau (at around 100 million articles estimates vary) not for an edit rate plateau and even there we would expect growth at a level roughly equal to web growth rates.
Yeah, cos the population of editors to write the English wikipedia is growing at the web growth rate due to the lack of censorship giving them impure thoughts.
And we clearly need one article for every few English speakers. In my view participation in the wikipedia needs to be enforced by law, like jury duty, upon pain of imprisonment and with absolutely no stupid restrictions like 'notability'.
Looks like the wikipedia is going to level off at just under 3.5 million articles to me. Something Must Be Done to write 97 million articles!
-- geni
2008/8/8 Ian Woollard ian.woollard@gmail.com:
2008/8/8 geni geniice@gmail.com: Yeah, cos the population of editors to write the English wikipedia is growing at the web growth rate due to the lack of censorship giving them impure thoughts.
Anything less than web growth rate shows we are being less successful than recruiting in the past.
Looks like the wikipedia is going to level off at just under 3.5 million articles to me. Something Must Be Done to write 97 million articles!
Actually writing them wouldn't for the most part be a problem (mostly once you deal with some of the copyright issues). Ensuring they are about things other than chemicals and astronomical objects may be more of a challenge.
On 08/08/2008, geni geniice@gmail.com wrote:
2008/8/8 Ian Woollard ian.woollard@gmail.com:
2008/8/8 geni geniice@gmail.com: Yeah, cos the population of editors to write the English wikipedia is growing at the web growth rate due to the lack of censorship giving them impure thoughts.
Anything less than web growth rate shows we are being less successful than recruiting in the past.
Is the problem recruitment or retention?
Recruitment, in my totallyunscientific opinion.
A year ago, a jazz band that won several international awards, that had an acclaimed album out (produced by a noted famous producer, no less) would have had an article on Wikipedia. Someone's first thought when coming across the band would have been "I'll see what I can find on Wikipedia", and then "Oh, hang on, no article, I'll go and do a stub".
[[Empirical (jazz band)]] is such a band. And it was recently prominently featured on BoingBoing. That alone would have been enough, in the past, to get a Wikipedia article.
As it was, I created their article myself. I know next to nothing about them except that I have their debut album and that I quite like it.
Just a year ago you could have gotten away with creating a one-paragraph stub. Any bets how long it would have taken for the article to be speedied if I hadn't provided two sources?
Any guesses as to why the article was not created sooner?
Michel Vuijlsteke
2008/8/9 James Farrar james.farrar@gmail.com
On 08/08/2008, geni geniice@gmail.com wrote:
2008/8/8 Ian Woollard ian.woollard@gmail.com:
2008/8/8 geni geniice@gmail.com: Yeah, cos the population of editors to write the English wikipedia is growing at the web growth rate due to the lack of censorship giving them impure thoughts.
Anything less than web growth rate shows we are being less successful than recruiting in the past.
Is the problem recruitment or retention?
2008/8/9 Michel Vuijlsteke wikipedia@zog.org:
Recruitment, in my totallyunscientific opinion.
A year ago, a jazz band that won several international awards, that had an acclaimed album out (produced by a noted famous producer, no less) would have had an article on Wikipedia. Someone's first thought when coming across the band would have been "I'll see what I can find on Wikipedia", and then "Oh, hang on, no article, I'll go and do a stub".
[[Empirical (jazz band)]] is such a band. And it was recently prominently featured on BoingBoing. That alone would have been enough, in the past, to get a Wikipedia article.
As it was, I created their article myself. I know next to nothing about them except that I have their debut album and that I quite like it.
Just a year ago you could have gotten away with creating a one-paragraph stub. Any bets how long it would have taken for the article to be speedied if I hadn't provided two sources?
Any guesses as to why the article was not created sooner?
Michel Vuijlsteke
The problem with that theory is that news users haven't been able to create new articles for some time.
2008/8/8 geni geniice@gmail.com:
2008/8/8 Ian Woollard ian.woollard@gmail.com:
2008/8/8 geni geniice@gmail.com: Yeah, cos the population of editors to write the English wikipedia is growing at the web growth rate due to the lack of censorship giving them impure thoughts.
Anything less than web growth rate shows we are being less successful than recruiting in the past.
I'm sorry, I simply don't agree with this idea.
There simply isn't an infinite amount of human knowledge, there is, by definition a finite amount. And there's a rather smaller amount of encyclopedic knowledge than that.
I consider an update to the wikipedia to be fixing a bug- we're adding something that isn't there already, that should be there. In software (which the wikipedia is really) fixing bugs is an exponential decay process.
The problem is, as the wikipedia is written, we get the low-hanging fruit early on, and then the remaining fruit is higher and higher up the tree of knowledge, and is harder to understand, less people have the knowledge, and the chances of somebody fixing its omission goes down.
We rode the curve of web growth well up to about late 2007, after that the curve of knowledge difficulty has had a more significant effect and can be expected to decrease the growth rate even further over time.
I see no evidence in the data for the idea that the wikipedia is causing the reduction in growth, and I see strong evidence that it is simply getting harder and harder to add new knowledge, because you have to spot a rare gap that hasn't already been filled.
Basically, the argument in:
http://en.wikipedia.org wiki/Wikipedia:Modelling_Wikipedia%27s_growth#Is_the_growth_in_article_count_of_Wikipedia_logistic.3F
is that the data is consistent with the growth of the wikipedia following a logistics curve, and I firmly believe that this is what is happening- there's two exponentials at work here.
You could, by all means add a whole bunch of largely useless information into the wikipedia, but that won't make the wikipedia materially more useful to people- so the *effective* size- the wikipedia that people actually look at, will continue to follow the curve that it's already on.
-- geni
2008/8/8 geni geniice@gmail.com:
2008/8/8 Ian Woollard ian.woollard@gmail.com:
2008/8/8 geni geniice@gmail.com: Yeah, cos the population of editors to write the English wikipedia is growing at the web growth rate due to the lack of censorship giving them impure thoughts.
Anything less than web growth rate shows we are being less successful than recruiting in the past.
I'm sorry, I simply don't agree with this idea.
There simply isn't an infinite amount of human knowledge, there is, by definition a finite amount. And there's a rather smaller amount of encyclopedic knowledge than that.
I consider an update to the wikipedia to be fixing a bug- we're adding something that isn't there already, that should be there. In software (which the wikipedia is really) fixing bugs is an exponential decay process.
The problem is, as the wikipedia is written, we get the low-hanging fruit early on, and then the remaining fruit is higher and higher up the tree of knowledge, and is harder to understand, less people have the knowledge, and the chances of somebody fixing its omission goes down.
I would go further than that, humans are quite limited in their interests and attention spans. Excellent coverage of highly significant issues is desirable, but very few people can keep track of more than a few hundred such issues.
Fred
On Fri, Aug 8, 2008 at 10:03 AM, David Gerard dgerard@gmail.com wrote:
Here's some numbers: http://oc-co.org/?p=124
Looks like they made a classic graphing error of not removing the the final partial datapoint causing the stupendous downtick at the end.