On 02/12/2012 01:31 PM, Pavel Tkachenko wrote:
Pavel, I asked Oliver Keyes, and he said that http://meta.wikimedia.org/wiki/Research:Newbie_reverts_and_article_length may be of interest. He's not on this list, so if you have thoughts about
...
On Sun, 12 Feb 2012 07:12:54 +0100, Mike Dupont jamesmikedupont@googlemail.com wrote:
I have stopped editing because of harassment and wiki-stalking from WhiteWriter and no one stops him from doing the same to many other people. There are gangs of radicals that live on wikipedia and just make life miserable for others, and no visual editor will stop that.
...which only seconds Yuri's social factor thought.
Still, Yuri, how do you oppose the WMF studies Oliver has presented earlier? About the factor of "any markup by default".
I do not oppose those studies at all, nor do I deny their integrity.
Just that the research isn't accounting for much of the important factors besides the plain volume of text. What would be those, and how those might be accounted for, I plainly don't know.
Just off the top of my head, no saying how it's useful: attempts to edit what long pages were more likely to be reverted? Were the long pages the same over the years? Did the contested pages actually grow? Who were the reverters?
Just like you say, one might form a distinct impression that the participation numbers indeed *are* falling. But will *enabling* the editing with a (rushed?) visual job actually *help*?
I'm totally not blocking the mouths of visual tool-hungry masses, but why must things be done in such fashion? There were those leisurely "new parser a year" years, and now, suddenly, there is a rush? A complex tool like that, operating on a content corpus which was created without it, is just bound to break things. Are there a projections for the breakage numbers and impact? I wouldn't say Wikipedia is fit for the UNESCO heritage (what about the conservation, eh?), but it is good for some uses, after all.
Now, why not follow the MoinMoin example, and construct an extension for loading the wiki-page into OpenOffice, benefiting from the fact that the return path is already well-covered? "Order of magnitude" more simple.
Yury