From: Edward H. Trager (ehtrager@umich.edu)
Date: Thu Oct 20 2005 - 12:43:37 CST
Hi, Markus,
On Thursday 2005.10.20 09:15:19 -0700, Markus Scherer wrote:
> On 10/20/05, Edward H. Trager <ehtrager@umich.edu> wrote:
> > ( Wikipedia is wide-open. It is completely *promiscuous*. So why is it growing so fast,
> > and why is the quality of the articles consistently so high? Why isn't it just
> > full of complete nonsense? : http://www.kuro5hin.org/story/2001/9/24/43858/2479 )
>
> Personally, I liked the few wikipedia articles that I looked up so
> far, but it's certainly vulnerable and apparently admittedly
> imperfect: See The Register "Wikipedia founder admits to serious
> quality problems"
> http://www.theregister.com/2005/10/18/wikipedia_quality_problem/
I am sure Wikipedia is not a magic bullet for solving human error and error correction
issues. Wikipedia is still a new and young system which does exhibit very high quality
articles in certain domains. At least I have found the articles on world languages, cultures, and
writing systems to be pretty good. Perhaps Wikipedia articles on more academic subjects
are of high quality, while articles more on the "People Magazine" side of the spectrum
are trashier and contain more factual errors? Nicholas Carr's selection of Bill Gates
and Jane Fonda as "random" entries (as reported in the Register article) is absurdly non-random!
It would be *much* better
to do a real statistical sampling of Wikipedia articles across different knowledge domains
and have experts in those knowledge domains independently score their perceptions of
article quality.
In my original posting, I had
considered also mentioning Open Source software development methodologies which,
viewed as organic systems, are also pretty good at finding and correcting errors on
timescales that are shorter than in traditional software development systems. Open Source
can also be considered a new and young system. Open Source is also not a perfect solution.
Perhaps it would have been better if I had used the MathWorld project (http://mathworld.wolfram.com/)
as an example instead of Wikipedia. MathWorld is certainly a Wikipedia-like project, but focused on
just one knowledge domain unlike Wikipedia which tries to cover all. It is probably true
that MathWorld articles are written primarily by people who are very serious about mathematics,
whether they be academics, graduate students, or "amateur" mathematicians. The seriousness of
the authors and the audience probably contributes to a high level of quality with few factual
errors and rapid correction of factual errors when they occur. These are desirable traits.
Suppose that there existed a "UnicodeWorld" project that looked something like the aforementioned
MathWorld. Such a site would be of primary interest to people who are very serious about languages,
scripts, linguistics, font technologies, internationalization, localization, and so on. In other
words, the people who read this list. As a result,
it is almost certain that a high level of quality would dominate and the error-correction rate would
be very rapid.
- Ed Trager
>
This archive was generated by hypermail 2.1.5 : Thu Oct 20 2005 - 12:36:52 CST