en:wp's new article rate is slowing down:
http://www.andrewlih.com/blog/2007/06/28/wikipedia-plateau/
- d.
On 6/29/07, David Gerard dgerard@gmail.com wrote:
en:wp's new article rate is slowing down:
http://www.andrewlih.com/blog/2007/06/28/wikipedia-plateau/
- d.
Andrew does make several very thoughtful arguments about the probable cause, but I think that the real reason for decline in the rate of growth is this:
http://lists.wikimedia.org/pipermail/wikien-l/2005-December/033880.html
From December 2005, which appears to be the inflection point of the
graph (exponential turns logistic).
--Gracenotes
Gracenotes wrote:
Andrew does make several very thoughtful arguments about the probable cause, but I think that the real reason for decline in the rate of growth is this:
http://lists.wikimedia.org/pipermail/wikien-l/2005-December/033880.html
A pity nobody had any plans to monitor and analyze the results of this "experiment" back when it was initiated. Or, for that matter, any time since then.
If experimentation was really the point, how about we finally re-enable anon article creation and see whether the article creation rate changes again?
On 6/30/07, Bryan Derksen bryan.derksen@shaw.ca wrote:
If experimentation was really the point, how about we finally re-enable anon article creation and see whether the article creation rate changes again?
I have no doubts that rate would change, but is not going to do article quality any good.
MacGyverMagic/Mgm wrote:
On 6/30/07, Bryan Derksen bryan.derksen@shaw.ca wrote:
If experimentation was really the point, how about we finally re-enable anon article creation and see whether the article creation rate changes again?
I have no doubts that rate would change, but is not going to do article quality any good.
You've got nothing to back that assertion up because _no analysis was done_ when anon article creation was disabled in the first place. How do you know whether article quality went up or down?
This has always been my main beef regarding that Siegenthaler reaction, it was supposedly done as an experiment but the most important part of experimentation was never even planned for. It seemed more like a knee-jerk PR move to me.
Bryan Derksen wrote:
MacGyverMagic/Mgm wrote:
On 6/30/07, Bryan Derksen bryan.derksen@shaw.ca wrote:
If experimentation was really the point, how about we finally re-enable anon article creation and see whether the article creation rate changes again?
I have no doubts that rate would change, but is not going to do article quality any good.
You've got nothing to back that assertion up because _no analysis was done_ when anon article creation was disabled in the first place. How do you know whether article quality went up or down?
This has always been my main beef regarding that Siegenthaler reaction, it was supposedly done as an experiment but the most important part of experimentation was never even planned for. It seemed more like a knee-jerk PR move to me.
WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: http://lists.wikimedia.org/mailman/listinfo/wikien-l
Do WP:AFC for a few days, keeping in mind that the submissions there are made right next to a set of -very specific instructions-. Then see if you think anon page creation would be a great thing. If you think CAT:CSD backlogs -now-...
Todd Allen wrote:
Do WP:AFC for a few days, keeping in mind that the submissions there are made right next to a set of -very specific instructions-. Then see if you think anon page creation would be a great thing. If you think CAT:CSD backlogs -now-...
But can you actually _show_ that the quality of new articles went up when anon article creation was disabled? Sampling of new articles beforehand wasn't done and AfC didn't exist before anon article creation was disabled so there's no data to compare the current situation to. I'm not going to make an argument based just on my subjective opinion.
That's my whole _point._ As an experiment it was a complete failure since it produced no usable data. Perhaps by reversing it and actually collecting data this time we'll be able to salvage something from it.
Bryan Derksen wrote:
Todd Allen wrote:
Do WP:AFC for a few days, keeping in mind that the submissions there are made right next to a set of -very specific instructions-. Then see if you think anon page creation would be a great thing. If you think CAT:CSD backlogs -now-...
But can you actually _show_ that the quality of new articles went up when anon article creation was disabled? Sampling of new articles beforehand wasn't done and AfC didn't exist before anon article creation was disabled so there's no data to compare the current situation to. I'm not going to make an argument based just on my subjective opinion.
That's my whole _point._ As an experiment it was a complete failure since it produced no usable data. Perhaps by reversing it and actually collecting data this time we'll be able to salvage something from it.
WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: http://lists.wikimedia.org/mailman/listinfo/wikien-l
That would have been great, but since we still have a mechanism for anonymous editors to submit articles, I think it's the case that we -can- gather useful data from that, in terms of "Roughly how many pages that an anon wants to create would be pages we would actually wish to have, and how many of them would we need to take time to get rid of?"
We could certainly look at how many AfC submissions are accepted, and how many declined. And if anything, that may be a -higher- percentage of good pages than simply turning anon-creation back on, since AfC contains specific instructions, whereas "You can create this page..." often seems to give the mistaken impression that there are no minimum requirements.
I'm all for anon editing. A lot of edits made by anons are good ones, even if some are vandalism (and not as high a fraction of anon edits are vandalism as people seem to think). On the other hand, we get a very high number of garbage pages even from new registered contributors (and that certainly can be proven with hard evidence, I think someone at one point looked at what percentage new pages wound up deleted). That data seems to indicate that the threshold to create a page is, if anything, too -low-, and that if we're going to experiment we should do it with more restrictions, not fewer.
On Fri, 29 Jun 2007 23:03:24 -0600, Bryan Derksen bryan.derksen@shaw.ca wrote:
If experimentation was really the point, how about we finally re-enable anon article creation and see whether the article creation rate changes again?
I absolutely guarantee it would rise steeply. As would the deletion rate. Visit [[Specal:Newpages]] and [[WP:RA]] to prove this for yourself.
Guy (JzG)
Guy Chapman aka JzG wrote:
On Fri, 29 Jun 2007 23:03:24 -0600, Bryan Derksen bryan.derksen@shaw.ca wrote:
If experimentation was really the point, how about we finally re-enable anon article creation and see whether the article creation rate changes again?
I absolutely guarantee it would rise steeply. As would the deletion rate.
I should have clarified that I was talking about _net_ article creation rate. That's what the article that started this thread off was analyzing. If deletion rises as steeply as creation does there'd be no net change in Wikipedia's growth rate.
Visit [[Specal:Newpages]] and [[WP:RA]] to prove this for yourself.
Can't prove anything with those because there's no before-and-after information.
On 6/30/07, Bryan Derksen bryan.derksen@shaw.ca wrote:
Gracenotes wrote:
Andrew does make several very thoughtful arguments about the probable cause, but I think that the real reason for decline in the rate of growth is this:
http://lists.wikimedia.org/pipermail/wikien-l/2005-December/033880.html
A pity nobody had any plans to monitor and analyze the results of this "experiment" back when it was initiated. Or, for that matter, any time since then.
If experimentation was really the point, how about we finally re-enable anon article creation and see whether the article creation rate changes again?
If only someone had asked that the "experiment" be a valid one when it was started.
Oh right, I did.
Grr.
I did a little bit of statistical analysis at the time, which was as outrageously unscientific as a good faith effort could possibly be, and should be taken as a studied vignette.
I found that net creation rate dropped by perhaps 20% right around the change... and that new account creation went up significantly. A quote: "4 days out (discounting anything speedied in the intervening 4 days), around 45% of new articles were created by anons. Of these, around 80% are usable articles (the other 20% end up being deleted or merged). Most are stubs, and most do not know/conform to style guidelines."
http://en.wikipedia.org/wiki/User:Sj/Newpages
Being an eventualist and index-lover, I think these sorts of stubs, with or without proper style, are good for the encyclopedia and great for the index. I also think that Articles for Creation is pretty much a waste of time and energy, and very confusing for new users; and that barring indication of substantial improvement we should also turn off the restriction to see what the change is like in the other direction.
Most of the commentary I've seen about how necessary or pointless this change is/was has been speculation supported by passion.
SJ
On 7/5/07, The Cunctator cunctator@gmail.com wrote:
On 6/30/07, Bryan Derksen bryan.derksen@shaw.ca wrote:
Gracenotes wrote:
Andrew does make several very thoughtful arguments about the probable cause, but I think that the real reason for decline in the rate of growth is this:
http://lists.wikimedia.org/pipermail/wikien-l/2005-December/033880.html
A pity nobody had any plans to monitor and analyze the results of this "experiment" back when it was initiated. Or, for that matter, any time since then.
If experimentation was really the point, how about we finally re-enable anon article creation and see whether the article creation rate changes again?
If only someone had asked that the "experiment" be a valid one when it was started.
Oh right, I did.
Grr. _______________________________________________ WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: http://lists.wikimedia.org/mailman/listinfo/wikien-l
David Gerard wrote:
en:wp's new article rate is slowing down:
http://www.andrewlih.com/blog/2007/06/28/wikipedia-plateau/
- d.
WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: http://lists.wikimedia.org/mailman/listinfo/wikien-l
I'd be more interested in seeing a graph of the rate of -editing-. Quite realistically, I'm glad to see page creation slowing down, a lot of what gets created anymore is...stretching, to put it as nicely as I can, and then just takes effort to get rid of.
On 6/30/07, David Gerard dgerard@gmail.com wrote:
en:wp's new article rate is slowing down:
Are there any tools monitoring the deletion rate? I see a lot of comments on talk pages saying that Afd has a 75% success rate; has there been a shift since September 2006 ?
-- John