[WikiEN-l] Experiment on new pages

Mon Dec 5 16:56:52 UTC 2005

Jimmy Wales wrote:

>Delirium wrote:
>  
>
>>So, more interesting would be to address the fundamental problem---that
>>is, mark revisions that have been reviewed by [n] people.
>>    
>>
>
>Totally.   I think we'll be experimenting a lot in coming months,
>especially in en where it seems to me that a lot of time-honored
>processes are starting to be overwhelmed by sheer volume.
>  
>
That makes sense, although I prefer to view it as incrementally 
improving what already works, rather than ditching processes and 
bringing in new stuff.  For example, "in the beginning", we just looked 
at every edit on recent changes by someone we didn't recognize.  Now 
there are far too many edits, so we need some better way of organizing 
them, and marking which edits have been checked and which still need to 
be looked at.  Not really a fundamental shift in how Wikipedia works, 
just adding some tools to help us deal with things.

Slightly more fundamental would be displaying to users that a particular 
revision has a certain level of confidence.  This is already implicit 
for experienced Wikipedians---I think most of us know almost immediately 
whether a particular article is trustworthy or not, based on patterns 
like what the prose looks like, how wikified it is, how many people have 
edited it, what subject area it's in, etc.  But formalizing that a bit 
and making it explicit for newcomers and casual readers can't hurt---we 
all know that an unwikified page just created by 1 person and not edited 
by anyone else should be read with a grain of salt, but we can point 
that out to others in some automated way.  Again, just kind of 
incrementally improving how things work.

For the most part, I think we're doing pretty well, and there are lots 
of good ways to incrementally improve what's already gong well, it's 
just that period media frenzies kind of catch us when we're not yet ready.

-Mark