Jimmy Wales wrote on Sun Nov 13 03:48:10 UTC 2005:

you took any random 20 articles and compared them to their predecessors 2 years ago, virtually _all_ articles would have improved.

I agree. However, to judge a reference work, do you go by the best, or the worst, or the average quality? I think most readers would judge us by their first few impressions. As soon as they stray away from the featured article on the front page, or happen to run into the random pages, their impression will be not favorable. That's what I tried to find to out with my twenty random pages.

...

A thought experiment: if we were the editorial committee of an encyclopedia to be written from scratch and were given Wikipedia's current content as a basis (but not the user base), what would we do? I guess we would put our energy into improving the material, i.e. rewriting/deleting/merging most of it. But we would not try to acquire more articles of that quality. (Ah wait, we ARE the editorial committee of an encyclopedia to be written from scratch...)

I think this is a brilliant observation. :-)

I absolutely do think that acquisition of huge numbers of additional stubs on increasingly narrow topics ought not to be a priority, and certainly ought not to be allowed to get in the way of quality improvement on existing articles.

But it IS getting in the way. I have recently spent several hours patrolling the newpages and recent changes, and that was a very sobering experience. While I have no statistics (would be an interesting topic, though), I estimate that more than ninety percent of the newpages by anonymous contributors are unsalveageable. Of the rest, a sound majority qualifies for merging or heavy reworking. Plus an unbelievable amount of vandalism. I spent most of my time writing {{test}}, {{test2}}, {{test3}}, and {{test4}} messages at high speed, deleting, reverting and occasionally blocking a vandal. I can't remember seeing more than a few new pages by anonymous editors that resmbled anything remotely interesting or well-written. There were also many pages by newly-logged-in users that were no better. To summarise, the signal-to-noise ratio on recent changes is depressingly low, and many qualified editors spend hours and hours trying to pick the few jewels out of mountains of dirt.

If we look at the continuum of quality ranging from the average anon's average "Phil is gay" contribution to well-researched work by well-informed people, it is reasonable to assume that even above the newpages level there is a great body of contributions with a negative benefit-versus-work-needed balance. In my opinion, it far too great.

(At the same time, of course, it's worth pointing out that there's an easy mental trap to fall into... assuming that time people are spending working on obscure fancruft could in any way be diverted into increasing the quality of other articles. That's probably not true.)

Yes. (Warning: heresy ahead) I would simply reply that we could do without these people. Along with the vandals and trolls and clueless kids, they waste the time and energy that we should spend improving our encyclopedia. They waste an _enormous_ amount of time and energy while not helping us much.

Another thought experiment: What if we shut down new page creation for a year? Simply declare 2006 the year of quality improvement and accept no new pages until 2007. We might lose may good editors. We might lose many _potential_ good editors. We would drop in the Alexa ranking. We might lose donations (although I doubt it). But we could start lifting the quality of the average article to where it belongs, and we would have a chance to lift our reputation beyond the "public toilet" image that has deservedly been bestowed upon us.

So much for the heresy. Now flame me. Kosebamse