[WikiEN-l] About the reliability of the Wikipedia process and content

[Daniel Mayer]

--- David Gerard <fun at thingy.apana.org.au> wrote:
> We went through this (you and I) on this very list just recently. You're
> still reasoning along the lines of:
> 
> 1. We must do something.
> 2. This is something.
> 3. Therefore, we must do this.

No, my reasoning is that there is a good deal of FUD about Wikipedia right now
and that some of it is *valid.* Just pretending that everything is fine as is
does not move us forward. 
 
> Not only is this logically fallacious, you haven't really proven 1. as yet.
> Note that de: Wikipedia recently beat two commercial encyclopedias in blind
> testing without imposing review boards.

I would like to see the same test done on the English version before I jump to
that conclusion. 

We also need a way to mark articles as being OK for a 1.0 version. We *can not*
just give out the most recent database dump. 

> I suspect review boards will not scale. What is the processing rate of a
> review board likely to be? How many articles are there again? How many
> being created each day?

There could be different levels of review; at the lowest level we would have
what we have now, on top of that could be a reader feedback mechanism that is
in beta, the top level would be a review board review. The strengths of each
should be leveraged on the other. Users should have a choice of which versions
to use. 

Scale issues are very important so we must develop ways to to quickly review
content. Only by starting on that road can we work out bottlenecks. Note: that
the featured article process selects the best of Wikipedia content and is
therefore slow. Review boards would not have 'best of' as part of its criteria
but would look at a short list of criteria before marking it as reviewed. Such
as:

1) Does it cover the basic aspects of the subject?
2) Are there any obvious errors of fact? 
3) Are alternate views given an appropriate amount of space and qualified?
4) Mechanics (spelling, garmmar, etc) 

> While reviewing content sounds like a good ide (to me too), there should be
> a way to do it wihtout seeming to repudiate the concept of the wiki. 

I agree. That is why Bryan and I have suggested that users be given the most
recent version by default with a link to the last reviewed version. 

> I'm
> not actually wedded to the idea myself, but repudiating it as you describe
> would be extremely jarring culturally and - and this is the important point
> - demotivating for the volunteers.

I AM NOT REPUDIATING WIKI!! I want to add better review mechanisms onto what we
are already doing (which is not pure wiki). When I say that wiki is a means to
an end, I am just pointing out that the product is more important than the
process. But if changes in the process harm the product, then those changes
will have to be modified. That is as true for wiki as it would be for any
review system. 

> If it can at all be done within the wiki process, it should be. de: is
> solid evidence it can be achieved within the wiki process; neither your
> assertions nor those of the Britannica editor are solid evidence it can't.

Again, we need a way to check content before publishing it in print or digital
media. We also need a way to give users the *option* to see those checked
versions so that they do not have to worry if some nut job distorted the facts
presented in the article version the user is reading. 

Studies that show the average quality of Wikipedia are better than other
encyclopedias will be *very* little comfort for those people who happen upon a
particular article in a sub-par state (however brief that may be). 

> (And I'm still seeking details of what de: did to get to that standard!
> Could someone on de: please tell us? cc'd to wikipedia-l for this purpose)

Yes - that is very important information to know before we proceed with
anything. 

-- mav


		
__________________________________ 
Do you Yahoo!? 
The all-new My Yahoo! - Get yours free! 
http://my.yahoo.com